# Change is coming to roofing. You can't stop it. Disclosure: Films and drawings on this page are AI-generated illustrations. They show a mood or an idea, not real events, products or people. The old photos are real archival photos from NASA and the Library of Congress, shown untouched. The scenario numbers (Mild, Moderate and Aggressive) come from Foundation's research model, not a forecast. Other numbers come from the sources named in each scene's "What this shows" card (the Census, BLS, USDA and others). Company figures are the companies' own reports. Foundation works with roofing businesses and has a commercial interest in this change. The paper and its model have not been peer reviewed by outside academics. Useful destination: See what changes for your role, choose to take part, and leave with a three-step plan: owners map their exit, office teams and crews share this with their owner. Value promise: One clear next step for your role Subject or offer: Foundation gives roofing owners, office teams and crews a way to take part: a free 30-minute Map My Exit call for owners, an investor waitlist, and the research paper with every number. ## A working explanation ## Shared opening ### Attention:Mystery — Change is coming to roofing. Here's what it means for you. [Our view] Picture: [Storm film: generated illustration · mood only] Question: Is change coming to roofing, whether we like it or not? Discovery: In our view, yes. Nobody can stop it. What matters is what you do next. Mechanism: A storm film that moves with your scroll: you push toward it yourself. Visitor action: Scroll into the storm. Consequence: You start with our view stated plainly, labelled as a view and not a forecast. Portable contribution: Starting point: change is coming to roofing (our view). Next: What the storm is made of. ### Attention:Mystery — AI is changing office work. Robots may change roof work too. [Our view] Picture: [Storm film: generated illustration · mood only] Question: What is the storm made of? Discovery: Two things: AI in the office now, and robots that may come to roof work later. Mechanism: The words ride over the same storm, one screen later. Visitor action: Keep scrolling. Consequence: You know which two changes the rest of the page follows. Portable contribution: Two changes: AI in office work, robots in roof work. Next: Which one is you? ### Blank card: Shared opening - Change is coming to roofing.: Starting point: change is coming to roofing (our view). - AI is changing office work.: Two changes: AI in office work, robots in roof work. ## Which one is you? You can take part in this. Or you can let it happen to you. Question: Which one is you? - Roofing business owners (You own or run a roofing company): What changes for my company? See "The owner's path". - Office teams (You work in a roofing office): What changes in the office? See "The office's path". - Roofers and crews (You work on roofs): What changes on the roof? See "The crew's path". ## Shared opening (continued) ### Interest:Education — This isn't the first time. This isn't the first time. 'Computer' used to be a job title: a person who did math by hand. [Official description · NASA History Office] When electronic computers came, many learned to program them. Today there are about 5.3 million computer and math jobs. [Official description · NASA History Office] [Measured · BLS OEWS May 2025] Archival photo: NACA / NASA Ames Research Center, ARC-1943-AAL-4961 (1943). NASA media: generally not subject to copyright in the United States. No NASA endorsement implied. Archival photo: NACA Langley / NASA, photo by Taub, LRC-1957-B701_P-05391 (1957). NASA media: generally not subject to copyright in the United States. No NASA endorsement implied. Question: Has a new tool taken over a whole kind of work before? Discovery: Yes. 'Computer' was once a person's job. Many of those people learned to program the machines. Mechanism: Two real NASA photos, full screen and untouched, with a slow camera push and a hard cut. Visitor action: Scroll through 1943 and 1957. Consequence: You see that a job can change completely and still lead to more work. Portable contribution: Seen: human computers became programmers. Next: Telephone operators. ### Interest:Education — Telephone operators. This is the U.S. Capitol switchboard in 1959. In 1970, about 420,000 Americans worked as telephone operators. [Measured · Census 1970] Today, fewer than 40,000 do. Machines took over connecting calls; people took over the conversation. [Measured · BLS OEWS May 2025] [Economist's summary · Bessen] Question: What happened to telephone operators when machines connected the calls? Discovery: Operator jobs fell from about 420,000 to under 40,000. Millions now work the conversation instead. Mechanism: The real 1959 photo first, untouched. Then a separate dark stage: 21 figures, 1 per 20,000 operators. 19 fade. Visitor action: Scroll from the photo to the row of figures. Consequence: You see the task move to machines and the people move to the conversation. Portable contribution: Seen: machines took the connecting; people took the conversation. Next: Bank tellers. - 3D scene (history-crowd): Telephone operators. 1 figure = 20,000 operators. Measured · Census 1970 (Series D351); BLS OEWS May 2025. Not the same people, and not a one-for-one swap. - telephone operators in 1970: 21 figures (about 420,000 telephone operators). - telephone and switchboard operators today: 2 figures (about 37,710 telephone and switchboard operators). - Faded figures show the count that went down. They are not people out of work. - Photo, shown untouched before the figures: Telephone operators at the U.S. Capitol switchboard, 1959 (Archival photo · Library of Congress, Prints & Photographs Division, U.S. News & World Report Collection, LC-DIG-ds-04925 (Marion S. Trikosko, 1959)). ### Interest:Education — Bank tellers. ATMs arrived in the 1970s. Bank teller jobs didn't disappear. [Economist's summary · Bessen] Archival photo: Library of Congress, LC-USZ62-23640 (1908). Library of Congress: no known restrictions on publication. Question: Did ATMs end bank teller jobs? Discovery: No. Teller jobs didn't disappear when ATMs arrived. Mechanism: A real 1908 photo of a bank floor, shown untouched. Visitor action: Keep scrolling. Consequence: You hold the question: what happened to the tellers? Portable contribution: Seen: ATMs arrived; tellers stayed. Next: What the numbers did. ### Interest:Education — Teller jobs doubled. Teller jobs roughly doubled, to about 556,000 by 2010. Fewer branches and online banking have since cut them to about 330,000. [Measured · Census 1970; BLS OES May 2010; BLS OEWS May 2025] Question: What happened to teller jobs after ATMs? Discovery: They roughly doubled by 2010. Fewer branches and online banking have cut them since. Mechanism: A hard cut to a generated ink-drawing film of a 1970s drive-up bank (a second car rolls up; it plays once and holds), labelled as not a historical image. Visitor action: Keep scrolling. Consequence: A new tool can grow a job before a later tool shrinks it. Portable contribution: Seen: teller jobs about doubled after ATMs, then fell with online banking. Next: Farms. ### Interest:Education — Farms: 100 workers. In 1900, 41 of every 100 American workers farmed. [Measured · USDA Economic Research Service] By 2000, about 2 did. Yet farms now grow nearly three times what they did in 1948. [Measured · USDA Economic Research Service] Question: How many American workers farmed in 1900, and how many by 2000? Discovery: 41 of every 100 in 1900. About 2 by 2000. Mechanism: Two generated ink-drawing films, cut hard from 1900 (horses at a walk) to 2000 (one combine at work); each plays once and holds. 100 figures empty the field and the wheat grows about three times taller. Visitor action: Scroll from 1900 to 2000. Consequence: You see a whole way of working shrink to a sliver. Portable contribution: Seen: farm work went from 41 of 100 workers to about 2. Next: What the farms grow now. - 3D scene (history-crowd): 100 American workers. 1 figure = 1 in 100 workers. Measured · USDA Economic Research Service. - American workers farmed in 1900: Farm work 41 of 100, Other work 59 of 100 (about 41 of every 100 American workers). - American workers farmed in 2000: Farm work 2 of 100, Other work 98 of 100 (about 1.9 of every 100 American workers); the wheat stands 3 times as tall. - 1900 counts the whole workforce; 2000 counts employed workers. - Farm output is measured against 1948. The wheat height is a picture of that, not a measurement. ### Interest:Surprise — Office work changes first. In our model, AI changes office work first. Records, estimates, orders and schedules move the most. [Foundation's model, not a forecast] Picture: [Office film: generated illustration] Question: Which work changes first? Discovery: Office work. In our model, records, estimates, orders and schedules move the most. Mechanism: A generated film of a roofing office desk at dawn. Visitor action: Keep scrolling. Consequence: You know the office feels it first, and that switching jobs is hard. Portable contribution: Seen: office work changes first in the model. Next: Switching jobs. ### Interest:Surprise — Switching jobs is hard. Some office workers may have to switch to jobs like electrician and nurse. That's in Anthropic's bigger-change scenarios, and Anthropic says switching is hard. [Anthropic's working paper] Picture: [Office film: generated illustration] Question: Can office workers just switch jobs? Discovery: In Anthropic's bigger-change scenarios some office workers may have to switch to jobs like electrician and nurse, and Anthropic says switching is hard. Mechanism: The empty office screen film, dimmed, behind Anthropic's view. Visitor action: Keep scrolling. Consequence: You know switching jobs is not an easy answer. Portable contribution: Seen: switching jobs is hard (Anthropic's view). Next: The trades. ### Interest:Education — Some may move into the trades. That takes time, and the work may change while they learn it. [The paper's view] Picture: [Rig film: generated illustration · practice rig] Question: Are the trades a safe place to go? Discovery: The trades are changing too. A move into the trades takes time. Mechanism: A generated film of a humanoid robot on a practice rig in a workshop, labelled as a practice rig. Visitor action: Keep scrolling. Consequence: You see why "just learn a trade" is not the whole answer. Portable contribution: Seen: the trades are changing too. Next: What robot companies say. ### Interest:Education — What companies say. Renovate Robotics says it built a robot that installs asphalt shingles on steep roofs. Pilots were planned with contractors. [Company's own report] Picture: [Rig film: generated illustration · no robot shown] The picture behind is a generated illustration, not Renovate's or Unitree's robot or rig. Question: What have robot companies actually reported? Discovery: A shingle-laying robot with planned pilots, and thousands of humanoid robots shipped. Neither shows a robot doing useful roof work on its own. Mechanism: One company line over a generated film of an empty practice rig (a slow push-in, no robot in it); three more reports in the tap card. Visitor action: Keep scrolling. Tap for three more reports. Consequence: You can tell a company's claim from a test of real work. Portable contribution: Seen: what robot companies say, and what that does not show. Next: What no robot has done yet. ### Interest:Surprise — No robot has been shown doing a whole roof on its own. There is no set date for robots roofing on their own. [The paper's reading] Picture: [Rig film: generated illustration · no robot shown] Question: Has any robot roofed a whole house on its own? Discovery: No. Shingle-only robots have been announced and demonstrated. There is no set date for robots roofing on their own. Mechanism: One screen of words over the empty practice rig film (its own slow push-in), dimmed; on a phone, the push-in only, framed on the empty rig and the workshop, back and forth. Visitor action: Keep scrolling. Consequence: You see the honest edge of the evidence before the model. Portable contribution: Seen: no robot has done a whole roof on its own. Next: Foundation's model, as 100 people. ### Interest:Surprise — The 100. Here are 100 office workers. At the start of our model, about 4 are out of work or retraining. [Foundation's model, not a forecast] Mild case: still about 4. Moderate case: about 5. [Foundation's model, not a forecast] In Aggressive, the hardest case in the model, 15 of 100 office workers are out of work or retraining. [Foundation's model, not a forecast] Each ghost is 1 in 100 office workers. Out of work or retraining, in the hardest case. [Foundation's model, not a forecast] Picture: [Storm film: generated illustration · mood only] Question: In Foundation's model, how many of 100 office workers end up out of work or retraining? Discovery: About 4 at the start. About 4 in Mild, 5 in Moderate and 15 in Aggressive, the hardest case in the model. Mechanism: 100 solid figures under the storm. Scroll through the cases: each new ghost, a hollow dashed outline, turns one at a time in the front rows. Then the camera stands face to face with one ghost. Visitor action: Scroll through the scenarios. Tap "Try another scenario" to replay one. Consequence: You see the model as people, not a line on a chart. Portable contribution: Seen: 15 of 100 office workers out of work or retraining in the Aggressive scenario (model, not a forecast). Next: One roof. - 3D scene (people-field): 100 office workers. Each figure is 1 of 100 office workers. - office workers out of work or retraining at the start of the model: 96 of 100 working, 4 of 100 out of work or retraining (ghost) (exact value: 3.8 per 100). - office workers out of work or retraining, Mild scenario, model year 6: 96 of 100 working, 4 of 100 out of work or retraining (ghost) (exact value: 3.9 per 100). - office workers out of work or retraining, Moderate scenario, model year 6: 95 of 100 working, 5 of 100 out of work or retraining (ghost) (exact value: 4.6 per 100). - office workers out of work or retraining, Aggressive scenario (the hardest), model year 6: 85 of 100 working, 15 of 100 out of work or retraining (ghost) (exact value: 15.1 per 100). - Road Mild: office workers out of work or retraining, Mild scenario, model year 6. - Road Moderate: office workers out of work or retraining, Moderate scenario, model year 6. - Road Aggressive (hardest): office workers out of work or retraining, Aggressive scenario (the hardest), model year 6. - Foundation's model, not a forecast ### Interest:Surprise — Now one roof. Now one roof: 100 hours of work. Each course is 1 hour, not 1 person. [Foundation's model, not a forecast] In the Moderate scenario, machines take 14 of every 100 hours. People still do most of them. [Foundation's model, not a forecast] In the hardest case, machines do 42 of every 100 hours. 42 hours, not 42 roofers. The crew stays on the roof. [Foundation's model, not a forecast] Power looms took over 98% of the work needed to weave each yard of cloth. Weaving jobs still grew. [Economist's summary · Bessen] In the model, more people look for roofing work. And each roof needs fewer hours from people. [Foundation's model, not a forecast] Picture: [Storm film: generated illustration · mood only] Question: In the model, how much roof work do machines take, and what happens to the crew? Discovery: About 2, 14 or 42 of every 100 hours (Mild, Moderate, Aggressive). They are hours, not roofers. The crew stays on the roof. Mechanism: A house whose front roof is 100 hours laid as shingle courses in eight task zones. Machine hours fill zone by zone while a crew of four stays tied off. Visitor action: Scroll. Tap a zone to see its share. Consequence: You see that machine hours are not missing roofers. Portable contribution: Seen: 42 of 100 roof hours done by machines in the Aggressive scenario, with the crew still on the roof (model, not a forecast). Next: Your choice. - 3D scene (roof-house): One roof, 100 hours of work. One roof is 100 hours of roof work, laid as shingle courses in 8 task zones: Tear-off 14, Deck repair 12, Underlayment 12, Flashing 12, Rig safe access 10, Move materials 12, Install covering 22, Inspect and hand over 6. Hours, not people. The crew stays on the roof. A crew of 4 stays on the roof in every scenario. - hours of roof work done by machines, Mild scenario: 2 of 100 hours done by machines (exact value 2.0). Tear-off: 0 of its 14 hours done by machines (2% of the zone, the authors' assumption). Deck repair: 0 of its 12 hours done by machines (0% of the zone, the authors' assumption). Underlayment: 0 of its 12 hours done by machines (2% of the zone, the authors' assumption). Flashing: 0 of its 12 hours done by machines (0% of the zone, the authors' assumption). Rig safe access: 0 of its 10 hours done by machines (0% of the zone, the authors' assumption). Move materials: 1 of its 12 hours done by machines (5% of the zone, the authors' assumption). Install covering: 1 of its 22 hours done by machines (4% of the zone, the authors' assumption). Inspect and hand over: 0 of its 6 hours done by machines (0% of the zone, the authors' assumption). - hours of roof work done by machines, Moderate scenario: 14 of 100 hours done by machines (exact value 13.5). Tear-off: 2 of its 14 hours done by machines (10% of the zone, the authors' assumption). Deck repair: 0 of its 12 hours done by machines (2% of the zone, the authors' assumption). Underlayment: 2 of its 12 hours done by machines (15% of the zone, the authors' assumption). Flashing: 0 of its 12 hours done by machines (1% of the zone, the authors' assumption). Rig safe access: 0 of its 10 hours done by machines (0% of the zone, the authors' assumption). Move materials: 4 of its 12 hours done by machines (30% of the zone, the authors' assumption). Install covering: 6 of its 22 hours done by machines (28% of the zone, the authors' assumption). Inspect and hand over: 0 of its 6 hours done by machines (3% of the zone, the authors' assumption). - hours of roof work done by machines, Aggressive scenario: 42 of 100 hours done by machines (exact value 42.3). Tear-off: 6 of its 14 hours done by machines (45% of the zone, the authors' assumption). Deck repair: 2 of its 12 hours done by machines (15% of the zone, the authors' assumption). Underlayment: 6 of its 12 hours done by machines (50% of the zone, the authors' assumption). Flashing: 1 of its 12 hours done by machines (10% of the zone, the authors' assumption). Rig safe access: 0 of its 10 hours done by machines (0% of the zone, the authors' assumption). Move materials: 10 of its 12 hours done by machines (80% of the zone, the authors' assumption). Install covering: 16 of its 22 hours done by machines (75% of the zone, the authors' assumption). Inspect and hand over: 1 of its 6 hours done by machines (15% of the zone, the authors' assumption). - The people walking into the yard show newcomers looking for roofing work. They are an illustration, not a count: the model's assumption is 3, 10 or 20 new people per 100 roofers (Mild, Moderate, Aggressive). - Foundation's model, not a forecast - Machine hours count human work replaced, not machine runtime. - Each zone is rounded so the zones add up to the total. ### Decision:Delight — Take part, or wait? Nobody can stop this. Not us, not anyone. [Our view] Take part, or wait? [Our view, not the model] Picture: [Sky film: generated illustration · mood only] Question: What do you do next? Discovery: Nobody can stop this. Taking part means learning the new tools early. Waiting doesn't stop the change. Mechanism: Drag or hold your own figure. It walks into the light, or it walks back as the storm rolls in. The model crowd stays the same either way. Visitor action: Drag or hold your figure, or press "Take part" or "Wait and see". Consequence: Your choice goes on your plan card. The model numbers do not change. Portable contribution: Your choice: take part, or wait and see (our view). Next: What changes for you. - 3D scene (choice-walk): Take part, or wait. 100 office workers in Foundation's model, Aggressive scenario: 85 working and 15 out of work or retraining in the model, the same in both outcomes. You stands one step outside the crowd and is not counted. - Take part: Taking part means learning the new tools early. And having a say in how they're used. - Wait and see: Waiting is a choice too. The change comes anyway. - Our view, not the model. The model's 15 don't change. What you do next does. ### Blank card: Shared opening (continued) - This isn't the first time.: Seen: human computers became programmers. - Telephone operators.: Seen: machines took the connecting; people took the conversation. - Bank tellers.: Seen: ATMs arrived; tellers stayed. - Teller jobs doubled.: Seen: teller jobs about doubled after ATMs, then fell with online banking. - Farms: 100 workers.: Seen: farm work went from 41 of 100 workers to about 2. - Office work changes first.: Seen: office work changes first in the model. - Switching jobs is hard.: Seen: switching jobs is hard (Anthropic's view). - Some may move into the trades.: Seen: the trades are changing too. - What companies say.: Seen: what robot companies say, and what that does not show. - No robot has been shown doing a whole roof on its own.: Seen: no robot has done a whole roof on its own. - The 100.: Seen: 15 of 100 office workers out of work or retraining in the Aggressive scenario (model, not a forecast). - Now one roof.: Seen: 42 of 100 roof hours done by machines in the Aggressive scenario, with the crew still on the roof (model, not a forecast). - Take part, or wait?: Your choice: take part, or wait and see (our view). Take part, or wait (Our view, not the model): Take part / Wait and see ## The owner's path ### Decision:Surprise — You run the company. You run the company. Here's what changes for you. Picture: [Owner film: generated illustration] Question: What changes for a roofing company owner? Discovery: Your office changes first. Weather still rules the roof. Your people decide how well it goes. Mechanism: A generated film: an owner's hands on the wheel at dawn, watching the yard. Visitor action: Keep scrolling. Consequence: You know this path is about your company. Portable contribution: Path: roofing business owner. Next: Your office. ### Decision:Education — Your office changes first. Take one office job: the production manager. Each card is 1 hour of its work. [Foundation's model, not a forecast] In the hardest case, AI agents do about 55 of its 100 hours. [Foundation's model, not a forecast] New work for people doesn't appear by itself. In the Aggressive scenario, the model adds no new teaching or improvement work. [Our view] [Foundation's model, not a forecast] Picture: [Office film: generated illustration] Question: How much of one office job could AI agents do in the model? Discovery: Take the production manager: about 11, 32 or 55 of every 100 hours (Mild, Moderate, Aggressive). In the Aggressive scenario the model adds no new work for people. Mechanism: The office wall: 100 hour-cards on a job board, grouped into the production manager's eight tasks. Cards turn teal as AI agents take hours; gold cards are new work for people. Visitor action: Scroll. Tap a task to read its share. Consequence: You know where the first hours move, and that new roles have to be built. Portable contribution: Seen: about 55 of the production manager's 100 hours done by AI agents in the Aggressive scenario (model, not a forecast). Next: What rain does to machines. - 3D scene (office-wall): One office job, 100 hours of work. One job is 100 office hours, pinned as cards in 8 tasks: Inspect the roof 12, Build the estimate 16, Order materials 12, Schedule the crew 15, Track costs and progress 12, Check safety and quality 15, Coordinate people 12, Handle surprises 6. Hours, not people. - office hours done by AI agents · Mild: 11 of 100 hours done by AI agents; 6 gold cards (New work for people). Inspect the roof: 1 of its 12 hours done by AI agents (5% of the task, the authors' assumption). Build the estimate: 3 of its 16 hours done by AI agents (20% of the task, the authors' assumption). Order materials: 2 of its 12 hours done by AI agents (15% of the task, the authors' assumption). Schedule the crew: 2 of its 15 hours done by AI agents (15% of the task, the authors' assumption). Track costs and progress: 3 of its 12 hours done by AI agents (25% of the task, the authors' assumption). Check safety and quality: 0 of its 15 hours done by AI agents (0% of the task, the authors' assumption). Coordinate people: 0 of its 12 hours done by AI agents (5% of the task, the authors' assumption). Handle surprises: 0 of its 6 hours done by AI agents (0% of the task, the authors' assumption). - office hours done by AI agents · Moderate: 32 of 100 hours done by AI agents; 8 gold cards (New work for people). Inspect the roof: 2 of its 12 hours done by AI agents (15% of the task, the authors' assumption). Build the estimate: 8 of its 16 hours done by AI agents (50% of the task, the authors' assumption). Order materials: 6 of its 12 hours done by AI agents (50% of the task, the authors' assumption). Schedule the crew: 6 of its 15 hours done by AI agents (40% of the task, the authors' assumption). Track costs and progress: 9 of its 12 hours done by AI agents (70% of the task, the authors' assumption). Check safety and quality: 0 of its 15 hours done by AI agents (2% of the task, the authors' assumption). Coordinate people: 1 of its 12 hours done by AI agents (10% of the task, the authors' assumption). Handle surprises: 0 of its 6 hours done by AI agents (2% of the task, the authors' assumption). - office hours done by AI agents · hardest case: 55 of 100 hours done by AI agents; 0 gold cards (New work for people). Inspect the roof: 4 of its 12 hours done by AI agents (35% of the task, the authors' assumption). Build the estimate: 13 of its 16 hours done by AI agents (80% of the task, the authors' assumption). Order materials: 10 of its 12 hours done by AI agents (85% of the task, the authors' assumption). Schedule the crew: 11 of its 15 hours done by AI agents (75% of the task, the authors' assumption). Track costs and progress: 11 of its 12 hours done by AI agents (95% of the task, the authors' assumption). Check safety and quality: 2 of its 15 hours done by AI agents (10% of the task, the authors' assumption). Coordinate people: 4 of its 12 hours done by AI agents (30% of the task, the authors' assumption). Handle surprises: 0 of its 6 hours done by AI agents (5% of the task, the authors' assumption). - Gold cards are hours of new work, not new hires. - Foundation's model, not a forecast - Task names come from O*NET task statements for construction managers and supervisors. - Support time still applies: the model adds 0.10 hours of support for every office hour AI agents take over. - Each task is rounded so the tasks add up to the total. ### Decision:Education — One question for every robot vendor. Ask every robot vendor one question: what does each accepted task cost? [Our view] Picture: [Imagined future. Generated illustration, not footage.] Question: Does a roofing robot work every hour of the year? Discovery: No. Roofers stop for bad weather and winter, and we assume machines would too. So judge a machine by what each accepted task costs. Mechanism: One line over an imagined-future film of a hauling machine sitting idle in heavy rain, with a slow camera push-in (a seamless loop; a phone gets its own tall version, the machine above the words); the weather and season arithmetic sits behind a tap. Visitor action: Keep scrolling. Tap for the season arithmetic. Consequence: You have one question to ask any robot vendor. Portable contribution: Ask every robot vendor: what does each accepted task cost? Next: What taking part looks like. ### Decision:Delight — Taking part, for owners. Let your crew help pick what changes. Train people before work moves. [Our view] Question: What does taking part look like for an owner? Discovery: Let your crew help pick what changes, and train people before work moves. Mechanism: A generated film of a training bay: an instructor guides a hauling machine up a practice ramp while trainees watch. Visitor action: Keep scrolling to your plan. Consequence: You leave with a first step for your company. Portable contribution: Plan: let my crew help pick the first task to test. Next: The future we can build. ### Blank card: My plan to take part (owner) Three steps, in our view. - You run the company.: Path: roofing business owner. - Your office changes first.: Seen: about 55 of the production manager's 100 hours done by AI agents in the Aggressive scenario (model, not a forecast). - One question for every robot vendor.: Ask every robot vendor: what does each accepted task cost? - Taking part, for owners.: Plan: let my crew help pick the first task to test. - The first task we would test: #### My plan to take part (our view) - [ ] Let my crew help pick the first task to test. - [ ] Ask every robot vendor: what does each accepted task cost? - [ ] Map My Exit: free 30-minute call for roofing owners. #### See the numbers The model's numbers are in the research paper: foundationproject.net/research/roofing-scenarios. Other sources are listed under Sources and limits. #### Foundation's interest Foundation works with roofing businesses. We have a commercial interest in this change. The paper and its model have not been peer reviewed by outside academics. ## The office's path ### Decision:Surprise — You keep the office running. You keep the office running. In our model, your work changes first. [Foundation's model, not a forecast] Picture: [Front-desk film: generated illustration · mood only] Question: What changes for people who work in a roofing office? Discovery: In the model your work changes first. Records, estimates, orders and schedules move most. Safety checks and surprises stay mostly with people. Mechanism: The office film again: the desk, the lamp, the coffee. Visitor action: Keep scrolling. Tap for each task. Consequence: You know which of your tasks move and which stay. Portable contribution: Seen: about 55 of the production manager's 100 hours done by AI agents in the Aggressive scenario (model, not a forecast). Next: Your week, in hours. ### Decision:Education — Your week, in hours. Take one office job: the production manager. Each card is 1 hour of its work. [Foundation's model, not a forecast] In the hardest case, AI agents do about 55 of its 100 hours. Records, estimates, orders and schedules move the most. [Foundation's model, not a forecast] Safety checks and surprises stay mostly with people. [Our view] Picture: [Office film: generated illustration] Question: How much of one office job could AI agents do in the model? Discovery: Take the production manager: about 11, 32 or 55 of every 100 hours (Mild, Moderate, Aggressive). Records, estimates, orders and schedules move the most; safety checks and surprises stay mostly with people. Mechanism: The office wall: 100 hour-cards on a job board, grouped into the production manager's eight tasks. Visitor action: Scroll. Tap a task to read its share. Consequence: You know which of your tasks move and which stay. Portable contribution: Seen: about 55 of the production manager's 100 hours done by AI agents in the Aggressive scenario (model, not a forecast). Next: The conversation. - 3D scene (office-wall): One office job, 100 hours of work. One job is 100 office hours, pinned as cards in 8 tasks: Inspect the roof 12, Build the estimate 16, Order materials 12, Schedule the crew 15, Track costs and progress 12, Check safety and quality 15, Coordinate people 12, Handle surprises 6. Hours, not people. - office hours done by AI agents · Mild: 11 of 100 hours done by AI agents; 6 gold cards (New work for people). Inspect the roof: 1 of its 12 hours done by AI agents (5% of the task, the authors' assumption). Build the estimate: 3 of its 16 hours done by AI agents (20% of the task, the authors' assumption). Order materials: 2 of its 12 hours done by AI agents (15% of the task, the authors' assumption). Schedule the crew: 2 of its 15 hours done by AI agents (15% of the task, the authors' assumption). Track costs and progress: 3 of its 12 hours done by AI agents (25% of the task, the authors' assumption). Check safety and quality: 0 of its 15 hours done by AI agents (0% of the task, the authors' assumption). Coordinate people: 0 of its 12 hours done by AI agents (5% of the task, the authors' assumption). Handle surprises: 0 of its 6 hours done by AI agents (0% of the task, the authors' assumption). - office hours done by AI agents · Moderate: 32 of 100 hours done by AI agents; 8 gold cards (New work for people). Inspect the roof: 2 of its 12 hours done by AI agents (15% of the task, the authors' assumption). Build the estimate: 8 of its 16 hours done by AI agents (50% of the task, the authors' assumption). Order materials: 6 of its 12 hours done by AI agents (50% of the task, the authors' assumption). Schedule the crew: 6 of its 15 hours done by AI agents (40% of the task, the authors' assumption). Track costs and progress: 9 of its 12 hours done by AI agents (70% of the task, the authors' assumption). Check safety and quality: 0 of its 15 hours done by AI agents (2% of the task, the authors' assumption). Coordinate people: 1 of its 12 hours done by AI agents (10% of the task, the authors' assumption). Handle surprises: 0 of its 6 hours done by AI agents (2% of the task, the authors' assumption). - office hours done by AI agents · hardest case: 55 of 100 hours done by AI agents; 0 gold cards (New work for people). Inspect the roof: 4 of its 12 hours done by AI agents (35% of the task, the authors' assumption). Build the estimate: 13 of its 16 hours done by AI agents (80% of the task, the authors' assumption). Order materials: 10 of its 12 hours done by AI agents (85% of the task, the authors' assumption). Schedule the crew: 11 of its 15 hours done by AI agents (75% of the task, the authors' assumption). Track costs and progress: 11 of its 12 hours done by AI agents (95% of the task, the authors' assumption). Check safety and quality: 2 of its 15 hours done by AI agents (10% of the task, the authors' assumption). Coordinate people: 4 of its 12 hours done by AI agents (30% of the task, the authors' assumption). Handle surprises: 0 of its 6 hours done by AI agents (5% of the task, the authors' assumption). - Gold cards are hours of new work, not new hires. - Foundation's model, not a forecast - Task names come from O*NET task statements for construction managers and supervisors. - Support time still applies: the model adds 0.10 hours of support for every office hour AI agents take over. - Each task is rounded so the tasks add up to the total. ### Decision:Education — The conversation stays. Machines took over connecting calls. People took over the conversation. [Economist's summary · Bessen] AI can draft the estimate. People still earn the customer's trust. [Our view] Picture: [Front-desk film: generated illustration · mood only] Question: What part of office work stays with people? Discovery: The conversation. Machines took over connecting calls; people took over the conversation. AI can draft the estimate; people earn the customer's trust. Mechanism: A generated film of a person on a headset at a roofing front desk, face out of frame. Visitor action: Keep scrolling. Consequence: You see where your value moves. Portable contribution: Seen: AI can draft the estimate; people still earn the customer's trust (our view). Next: What taking part looks like. ### Decision:Delight — Taking part, for office teams. Learn to run the AI tools and check their work. Own the customer conversation. [Our view] Question: What does taking part look like in the office? Discovery: Learn to run the AI tools and check their work. Own the customer conversation. Mechanism: The office film's last seconds, held. Visitor action: Keep scrolling to your plan. Consequence: You leave with a first step, and a reason to share this with your owner. Portable contribution: Plan: learn to run the AI tools and check their work. Next: The future we can build. ### Blank card: My plan to take part (office) Three steps, in our view. - You keep the office running.: Seen: about 55 of the production manager's 100 hours done by AI agents in the Aggressive scenario (model, not a forecast). - Your week, in hours.: Seen: about 55 of the production manager's 100 hours done by AI agents in the Aggressive scenario (model, not a forecast). - The conversation stays.: Seen: AI can draft the estimate; people still earn the customer's trust (our view). - Taking part, for office teams.: Plan: learn to run the AI tools and check their work. - The AI tool I will learn first: #### My plan to take part (our view) - [ ] Learn to run the AI tools and check their work. - [ ] Own the customer conversation. - [ ] Share this with my owner. #### See the numbers The model's numbers are in the research paper: foundationproject.net/research/roofing-scenarios. Other sources are listed under Sources and limits. #### Foundation's interest Foundation works with roofing businesses. We have a commercial interest in this change. The paper and its model have not been peer reviewed by outside academics. ## The crew's path ### Decision:Surprise — You work on the roof. You work on the roof. Here's what changes for you. Picture: [Crew film: generated illustration · mood only] Question: What changes for roofers and crews? Discovery: Some roof hours may move to machines. Your skill is in the hours that stay. Mechanism: A generated film of roofers tied off to ridge anchors at first light. Visitor action: Keep scrolling. Consequence: You know this path is about your hours on the roof. Portable contribution: Path: roofer or crew. Next: Which of your hours could change. ### Decision:Education — Which of your hours could change. In the model, machines take the most hours from carrying materials and laying shingles. Rigging safe access stays with people in every scenario. [Foundation's model, not a forecast] These are hours, not people. Your skill is in the hours that stay. [Our view] Picture: [Sky film: generated illustration · mood only] Question: Which roof tasks lose the most hours to machines in the model? Discovery: Carrying materials and laying shingles. Rigging safe access stays with people in every scenario. Mechanism: The house from before, close on the roof zones, with the crew still tied off. Visitor action: Scroll. Tap a zone to see its share. Consequence: You see which of your tasks hold their hours. Portable contribution: Seen: rigging safe access stays with people in every scenario (model, not a forecast). Next: What machines would need to learn. - 3D scene (roof-house): Your roof, 100 hours of work. One roof is 100 hours of roof work, laid as shingle courses in 8 task zones: Tear-off 14, Deck repair 12, Underlayment 12, Flashing 12, Rig safe access 10, Move materials 12, Install covering 22, Inspect and hand over 6. Hours, not people. Your skill is in the hours that stay. A crew of 4 stays on the roof in every scenario. - hours of roof work done by machines, Mild scenario: 2 of 100 hours done by machines (exact value 2.0). Tear-off: 0 of its 14 hours done by machines (2% of the zone, the authors' assumption). Deck repair: 0 of its 12 hours done by machines (0% of the zone, the authors' assumption). Underlayment: 0 of its 12 hours done by machines (2% of the zone, the authors' assumption). Flashing: 0 of its 12 hours done by machines (0% of the zone, the authors' assumption). Rig safe access: 0 of its 10 hours done by machines (0% of the zone, the authors' assumption). Move materials: 1 of its 12 hours done by machines (5% of the zone, the authors' assumption). Install covering: 1 of its 22 hours done by machines (4% of the zone, the authors' assumption). Inspect and hand over: 0 of its 6 hours done by machines (0% of the zone, the authors' assumption). - hours of roof work done by machines, Moderate scenario: 14 of 100 hours done by machines (exact value 13.5). Tear-off: 2 of its 14 hours done by machines (10% of the zone, the authors' assumption). Deck repair: 0 of its 12 hours done by machines (2% of the zone, the authors' assumption). Underlayment: 2 of its 12 hours done by machines (15% of the zone, the authors' assumption). Flashing: 0 of its 12 hours done by machines (1% of the zone, the authors' assumption). Rig safe access: 0 of its 10 hours done by machines (0% of the zone, the authors' assumption). Move materials: 4 of its 12 hours done by machines (30% of the zone, the authors' assumption). Install covering: 6 of its 22 hours done by machines (28% of the zone, the authors' assumption). Inspect and hand over: 0 of its 6 hours done by machines (3% of the zone, the authors' assumption). - hours of roof work done by machines, Aggressive scenario: 42 of 100 hours done by machines (exact value 42.3). Tear-off: 6 of its 14 hours done by machines (45% of the zone, the authors' assumption). Deck repair: 2 of its 12 hours done by machines (15% of the zone, the authors' assumption). Underlayment: 6 of its 12 hours done by machines (50% of the zone, the authors' assumption). Flashing: 1 of its 12 hours done by machines (10% of the zone, the authors' assumption). Rig safe access: 0 of its 10 hours done by machines (0% of the zone, the authors' assumption). Move materials: 10 of its 12 hours done by machines (80% of the zone, the authors' assumption). Install covering: 16 of its 22 hours done by machines (75% of the zone, the authors' assumption). Inspect and hand over: 1 of its 6 hours done by machines (15% of the zone, the authors' assumption). - Foundation's model, not a forecast - Five of the eight task names come from O*NET task statements for roofers; three are the authors' own. - Each zone is rounded so the zones add up to the total. ### Decision:Surprise — Your skill is what machines need to learn. If machines ever learn roof work, we think they'd learn from recordings of roofers. [Our view] Figure says it has paid $15M to people who record everyday tasks for it. It names homes, logistics centers, restaurants, factories and offices, not roofing. [Company's own report] Ask who owns the recordings of your work. We think workers should have a meaningful part in the gains created with their experience and effort. [Our view] Picture: [Capture film: generated illustration · practice rig] The film is a generated illustration, not Figure's recordings or anyone's. Question: What would a machine need from roofers to learn roof work? Discovery: Recordings of skilled work. So ask who owns the recordings of yours. Mechanism: A generated film of a roofer in a capture kit on a practice rig, cropped to the hands and the shingles. Visitor action: Keep scrolling. Consequence: You see that your skill has value, and one question to ask about it. Portable contribution: Plan: ask who owns the recordings of my work. Next: What taking part looks like. ### Decision:Delight — Taking part, for crews. Learn to run the machines before they show up. Teach the next roofer. [Our view] Question: What does taking part look like for a roofer? Discovery: Learn to run the machines before they show up. Teach the next roofer. Mechanism: A generated film of a training bay: an instructor guides a hauling machine up a practice ramp while trainees watch. Visitor action: Keep scrolling to your plan. Consequence: You leave with a first step, and a reason to share this with your owner. Portable contribution: Plan: learn the machines before they show up; teach the next roofer. Next: The future we can build. ### Blank card: My plan to take part (crew) Three steps, in our view. - You work on the roof.: Path: roofer or crew. - Which of your hours could change.: Seen: rigging safe access stays with people in every scenario (model, not a forecast). - Your skill is what machines need to learn.: Plan: ask who owns the recordings of my work. - Taking part, for crews.: Plan: learn the machines before they show up; teach the next roofer. - One thing I could teach a new roofer: #### My plan to take part (our view) - [ ] Learn the machines before they show up. - [ ] Teach the next roofer. - [ ] Ask who owns the recordings of my work. #### See the numbers The model's numbers are in the research paper: foundationproject.net/research/roofing-scenarios. Other sources are listed under Sources and limits. #### Foundation's interest Foundation works with roofing businesses. We have a commercial interest in this change. The paper and its model have not been peer reviewed by outside academics. ## Shared close ### Action:Education — Roof materials are changing. So are the standards. FORTIFIED, a stronger roof standard, passed 100,000 designations in 2026. [Program's own report · IBHS] Picture: [Materials film: generated illustration · mood only] The picture behind is a generated illustration, not a FORTIFIED roof or test. Question: Is roofing itself standing still? Discovery: No. FORTIFIED, a stronger roof standard, passed 100,000 designations in 2026, more than half in the last three years. Mechanism: A generated macro film across new roof surfaces at sunrise; the standard's numbers are its own report. Visitor action: Keep scrolling. Tap for the Hurricane Sally study. Consequence: You see roofing already changing for the better. Portable contribution: Seen: stronger roof standards are spreading. Next: An imagined future. ### Action:Delight — The machine carries. People lead. Roofs, materials and tools will change. The people who take part get to shape how. [Our view] Better tools let you take on bigger jobs. The machine carries the load. Your crew runs the job. [Our view] Picture: [Imagined future. Generated illustration, not footage.] Question: What could taking part build? Discovery: Roofs, materials and tools will change. The people who take part get to shape how, and can take on bigger jobs. Mechanism: An imagined future, generated: a low wheeled machine brings shingles to the hoist while two roofers, tied off, lead from the ridge. Visitor action: Keep scrolling. Consequence: You see a future where machines carry and people lead. Portable contribution: Seen: the people who take part get to shape how (our view). Next: Sunrise. ### Action:Delight — Roofers, then and now. In 1900, about 9,000 Americans worked as roofers. [Measured · Census 1900] By 1970, about 65,000. [Measured · Census 1970] You're out in front. Today there are about 167,000 roofer jobs. [Measured · BLS 2025] Picture: [Imagined future. Generated illustration, not footage.] The street is an imagined future, not a photo of today's roofers. Question: Has roofing work grown or shrunk over a century of new tools? Discovery: It grew: about 9,000 roofers in 1900, 65,000 in 1970 and about 167,000 jobs today. Mechanism: Roofer figures fill the sunrise street, 1 figure per 1,000 roofers, walking toward you. Your figure stands two body-lengths in front, not counted. Visitor action: Scroll from 1900 to today. Consequence: You end on growth, with your figure out in front. Portable contribution: Seen: roofing grew from about 9,000 roofers in 1900 to about 167,000 jobs today. Next: Your next step. - 3D scene (sunrise-crowd): Roofers in America. 1 figure = 1,000 roofers. Measured · Census 1900, 1970; BLS 2025. Census counted people, BLS counts jobs, and the country grew too. - roofers in 1900: 9 figures (about 9,000 roofers). - roofers in 1970: 65 figures (about 65,000 roofers). - roofer jobs today: 167 figures (about 166,900 roofers). - You stands in front of the crowd and is not counted. ### Action:Delight — Don't put your head in the sand. Take part. Foundation gives roofing owners, office teams and crews a way to take part. [Our view] Picture: [Imagined future. Generated illustration, not footage.] Question: What is your next step? Discovery: Owners: map your exit in a free 30-minute call. Office teams and crews: share this with your owner. Mechanism: The calls to action ride over the sunrise street, your role first. Visitor action: Pick your next step. Consequence: You leave with one clear next step for your role. Portable contribution: Next step: Map My Exit, or share this with my owner. Next: Your plan card is below. - Calls to action: - Roofing business owners: Map My Exit (Free 30-minute call for roofing owners): https://foundationproject.net/schedule; Make my plan (opens the plan card) - Office teams: Share this with your owner (shares the page on the Roofing business owners path); Map My Exit (For your owner: a free 30-minute call): https://foundationproject.net/schedule; Make my plan (opens the plan card) - Roofers and crews: Share this with your owner (shares the page on the Roofing business owners path); Map My Exit (For your owner: a free 30-minute call): https://foundationproject.net/schedule; Make my plan (opens the plan card) - Investors: Join the investor waitlist: https://foundationproject.net/investor-packet - See the numbers in the research paper: https://foundationproject.net/research/roofing-scenarios - Disclosure: Foundation works with roofing businesses. We have a commercial interest in this change. The paper and its model have not been peer reviewed by outside academics. ### Blank card: Shared close - Roof materials are changing. So are the standards.: Seen: stronger roof standards are spreading. - The machine carries. People lead.: Seen: the people who take part get to shape how (our view). - Roofers, then and now.: Seen: roofing grew from about 9,000 roofers in 1900 to about 167,000 jobs today. - Don't put your head in the sand. Take part.: Next step: Map My Exit, or share this with my owner. ## Claims, sources and limits Review is an editorial status. A reviewed claim is not independent proof of a measured outcome. ### h-computer-job / documented / reviewed Label: Official description 'Computer' used to be a job title. It meant a person who did math by hand. Scope: Human computers at NACA's Langley laboratory, from 1935 Limitations: NASA History Office article about Langley. It describes the job; it does not count human computers nationwide. - NASA History - When the Computer Wore a Skirt / line 689 - Selected supporting excerpt: Before the development of electronic computers, the term “computer” referred to people, not machines. It was a job title, designating someone who performed mathematical equations and calculations by hand. - Limits: NASA History Office web article, saved 2026-10-05. ### h-computers-400 / documented / reviewed Label: Official description By 1946, one lab alone had trained about 400 of them. Scope: Langley's computing department (NACA), 1935 to 1946 Limitations: One laboratory. The source says the department had trained about 400 women; it is not a national count. - NASA History - When the Computer Wore a Skirt / line 689 - Selected supporting excerpt: By 1946, as the overall supervisor for Computing, Tucker presided over a vastly expanded department that had trained about 400 women and placed them in sections across the facility. - Limits: NASA History Office web article, saved 2026-10-05. ### h-computers-program / documented / reviewed Label: Official description When electronic computers came, many learned to program them. Scope: Langley's human computers, from 1947 Limitations: NASA says human computers "took on programming duties"; it does not say how many did. The Vaughan page says many former West Computers joined the new computing division. - NASA History - When the Computer Wore a Skirt / line 729 - Selected supporting excerpt: With the introduction of the first electronic computers, human computers took on programming duties as well. - Limits: NASA History Office web article, saved 2026-10-05. - NASA - Dorothy Vaughan (people page) / line 709 - Selected supporting excerpt: Dorothy Vaughan and many of the former West Computers joined the new Analysis and Computation Division (ACD), a racially and gender-integrated group on the frontier of electronic computing. - Limits: NASA's biography page, saved 2026-10-05. ### h-vaughan / documented / reviewed Label: Official description Dorothy Vaughan and many former human computers joined NASA's new computing division. She became an expert FORTRAN programmer. Scope: Dorothy Vaughan, NACA and NASA Langley, from 1958 Limitations: NASA's own biography. The source names the former West Computers (Langley's segregated computing unit) and the Analysis and Computation Division. Text only on the page: no likeness, no AI. - NASA - Dorothy Vaughan (people page) / line 709 - Selected supporting excerpt: In 1958, when the NACA made the transition to NASA, segregated facilities, including the West Computing office, were abolished. - Limits: NASA's biography page, saved 2026-10-05. - NASA - Dorothy Vaughan (people page) / line 709 - Selected supporting excerpt: Dorothy Vaughan and many of the former West Computers joined the new Analysis and Computation Division (ACD) - Limits: NASA's biography page, saved 2026-10-05. - NASA - Dorothy Vaughan (people page) / line 709 - Selected supporting excerpt: Dorothy Vaughan became an expert FORTRAN programmer - Limits: NASA's biography page, saved 2026-10-05. ### h-computer-growth / measured / reviewed Label: Official estimate Today there are about 5.3 million computer and math jobs in the U.S. In 1960 there were about 12,000 computer specialists, and about 2.5 million by 2000. Scope: U.S. computer specialists, 1960 and 2000 (BLS Monthly Labor Review, 2006); computer and mathematical occupations, May 2025 (BLS OEWS) Limitations: A wider group than the 1960 count: 2025 covers all computer and mathematical occupations (5,260,110 jobs). OEWS counts wage-and-salary jobs in the U.S. and leaves out the self-employed, so "Americans" means jobs here. - BLS Monthly Labor Review 2006 - Wyatt and Hecker, Occupational changes during the 20th century (Wayback) / lines 806–809 - Selected supporting excerpt: Computer specialists grew 95 times as a proportion of total employment between 1960 and 2000, from 0.02 percent to 1.92 percent. (See chart 3, top panel.) Employment grew from 12,000 to 2,496,000. - Limits: BLS Monthly Labor Review, March 2006 (Wyatt and Hecker), Wayback copy. Based on decennial Census data. - BLS OEWS May 2025 - Computer and mathematical occupations (API) / line 11 - Selected supporting excerpt: "seriesID": "OEUN000000000000015000001", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Computer and mathematical occupations (API) / line 14 - Selected supporting excerpt: "year": "2025", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Computer and mathematical occupations (API) / line 18 - Selected supporting excerpt: "value": "5260110" - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. ### h-operators / measured / reviewed Label: Official estimate In 1970, about 420,000 Americans worked as telephone operators. Today, fewer than 40,000 do. Scope: U.S. telephone operators: Census 1970 (Historical Statistics, Series D351) and BLS OEWS May 2025 (telephone operators plus switchboard operators) Limitations: 1970 counts people (420,000, aged 16 and over, in the economically active population). 2025 counts wage-and-salary jobs: 3,430 telephone operators plus 34,280 switchboard operators, including answering service, = 37,710. Not the same people, and not a one-for-one swap. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1687 - Selected supporting excerpt: Series D 233-682. Detailed Occupation of the Economically Active Population: 1900 to 197 0-—-Con. - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1688 - Selected supporting excerpt: {In thousands of persons 14 years old and over, except a8 indicated] - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1692 - Selected supporting excerpt: No. 16years|1l4 years; 1970 1960 1960 1950 1940 1930 1920 1910 1900 - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1771 - Selected supporting excerpt: 351 | Telephone operators_.---.--~--------------e- nee 420 421 372 372 367 875 214 249 190 98 19 - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. Row D351: 420 is the 1970 value for ages 16 and over (OCR reads the 1950 value 375 as 875). - BLS OEWS May 2025 - Telephone operators (API) / line 11 - Selected supporting excerpt: "seriesID": "OEUN000000000000043202101", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Telephone operators (API) / line 14 - Selected supporting excerpt: "year": "2025", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Telephone operators (API) / line 18 - Selected supporting excerpt: "value": "3430" - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Switchboard operators (API) / line 11 - Selected supporting excerpt: "seriesID": "OEUN000000000000043201101", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Switchboard operators (API) / line 14 - Selected supporting excerpt: "year": "2025", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Switchboard operators (API) / line 18 - Selected supporting excerpt: "value": "34280" - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. ### h-operators-row / inferred / reviewed Label: Derived arithmetic On the dark stage, 1 figure stands for 20,000 operators: 21 figures for 1970 and 2 for today. Scope: Scale of the row of figures (switchboards) Limitations: Our arithmetic: 420,000 ÷ 20,000 = 21; 37,710 ÷ 20,000 = 1.9, shown as 2. The figures fade; they are not shown as people put out of work. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1771 - Selected supporting excerpt: 351 | Telephone operators_.---.--~--------------e- nee 420 421 372 372 367 875 214 249 190 98 19 - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - BLS OEWS May 2025 - Telephone operators (API) / line 18 - Selected supporting excerpt: "value": "3430" - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Switchboard operators (API) / line 18 - Selected supporting excerpt: "value": "34280" - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. ### h-switch-to-desk / documented / reviewed Label: Attribution Machines took over connecting calls. People took over the conversation. Scope: Telephone operators and receptionists in the U.S., 20th century Limitations: Economist's summary: James Bessen (IMF, 2015) writes that automated phone systems meant fewer switchboard operators but more receptionists, who took over the human interaction tasks. BLS (2006) says automated switching affected telephone operators and that receptionists grew as a share of jobs through 2000. Not the same people, and not a one-for-one swap. - IMF F&D 2015 - Bessen, Toil and Technology (Wayback) / line 77 - Selected supporting excerpt: automated company phone systems meant fewer switchboard operators but more receptionists who took over the human interaction tasks switchboard operators previously performed. - Limits: James Bessen's article in the IMF's Finance & Development (March 2015), Wayback copy. An economist's summary, not data we checked. - BLS Monthly Labor Review 2006 - Wyatt and Hecker, Occupational changes during the 20th century (Wayback) / lines 1395–1396 - Selected supporting excerpt: Automated switching and voice messaging affected telephone operators; - Limits: BLS Monthly Labor Review, March 2006 (Wyatt and Hecker), Wayback copy. Based on decennial Census data. - BLS Monthly Labor Review 2006 - Wyatt and Hecker, Occupational changes during the 20th century (Wayback) / lines 1405–1406 - Selected supporting excerpt: and receptionists, increased as a percent of employment through 2000. - Limits: BLS Monthly Labor Review, March 2006 (Wyatt and Hecker), Wayback copy. Based on decennial Census data. ### h-desk-jobs / measured / reviewed Label: Official estimate About 910,000 receptionists and 2.6 million customer service reps work today. Scope: U.S. receptionists and information clerks (43-4171) and customer service representatives (43-4051), May 2025 Limitations: BLS OEWS wage-and-salary jobs: 910,180 and 2,595,750. "Receptionists" on the page is the OEWS group receptionists and information clerks. These jobs did not all come from operator work. - BLS OEWS May 2025 - Receptionists and information clerks (API) / line 11 - Selected supporting excerpt: "seriesID": "OEUN000000000000043417101", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Receptionists and information clerks (API) / line 14 - Selected supporting excerpt: "year": "2025", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Receptionists and information clerks (API) / line 18 - Selected supporting excerpt: "value": "910180" - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Customer service representatives (API) / line 11 - Selected supporting excerpt: "seriesID": "OEUN000000000000043405101", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Customer service representatives (API) / line 14 - Selected supporting excerpt: "year": "2025", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Customer service representatives (API) / line 18 - Selected supporting excerpt: "value": "2595750" - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. ### h-tellers / measured / reviewed Label: Official estimate + Attribution ATMs arrived in the 1970s. Bank teller jobs didn't disappear. They roughly doubled, to about 556,000 by 2010. Online banking has since cut them to about 330,000. Scope: U.S. bank tellers: Census 1970 (Series D355), BLS OES May 2010, BLS OEWS May 2025 Limitations: 1970 counts people (about 253,000, aged 16 and over); 2010 and 2025 count wage-and-salary jobs (556,310 and 329,480). "Roughly doubled" compares 1970 with 2010 (about 2.2 times). The ATM dates and "did not decrease" are Bessen's summary. BLS links the later fall to fewer branches and online banking; it does not put all of it down to online banking. - IMF F&D 2015 - Bessen, Toil and Technology (Wayback) / line 67 - Selected supporting excerpt: Automated teller machines (ATMs) were first installed in the United States and other developed economies in the 1970s. - Limits: James Bessen's article in the IMF's Finance & Development (March 2015), Wayback copy. An economist's summary, not data we checked. - IMF F&D 2015 - Bessen, Toil and Technology (Wayback) / line 69 - Selected supporting excerpt: in fact the number of bank teller jobs did not decrease as the ATMs were rolled out (see Chart 1). - Limits: James Bessen's article in the IMF's Finance & Development (March 2015), Wayback copy. An economist's summary, not data we checked. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1688 - Selected supporting excerpt: {In thousands of persons 14 years old and over, except a8 indicated] - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1692 - Selected supporting excerpt: No. 16years|1l4 years; 1970 1960 1960 1950 1940 1930 1920 1910 1900 - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1775 - Selected supporting excerpt: 355 | Bank tellers.............__---------- eee nen e eee eee 253 254 185 181 65) 1,681) 2-323 654 235 - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. Row D355: 253 is the 1970 value for ages 16 and over; the rest of this OCR row is garbled and is not used. - BLS OES May 2010 - Tellers (Wayback) / line 436 - Selected supporting excerpt: Occupational Employment and Wages, May 2010 - Limits: BLS Occupational Employment Statistics, May 2010, Wayback copy. Counts wage-and-salary jobs. - BLS OES May 2010 - Tellers (Wayback) / line 438 - Selected supporting excerpt: 43-3071 Tellers - Limits: BLS Occupational Employment Statistics, May 2010, Wayback copy. Counts wage-and-salary jobs. - BLS OES May 2010 - Tellers (Wayback) / line 468 - Selected supporting excerpt: 556,310 - Limits: BLS Occupational Employment Statistics, May 2010, Wayback copy. Counts wage-and-salary jobs. - BLS OEWS May 2025 - Tellers (API) / line 11 - Selected supporting excerpt: "seriesID": "OEUN000000000000043307101", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Tellers (API) / line 14 - Selected supporting excerpt: "year": "2025", - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS OEWS May 2025 - Tellers (API) / line 18 - Selected supporting excerpt: "value": "329480" - Limits: BLS public API, OEWS national employment series (annual 2025 estimate, May 2025 reference). Counts wage-and-salary jobs; leaves out the self-employed. Saved 2026-10-05. - BLS Occupational Outlook Handbook - Tellers (Wayback capture 2026-09-13) / line 635 - Selected supporting excerpt: the number of bank branches has been in decline due to technological change. As more people use online banking tools, such as mobile check deposits, fewer bank customers will visit the teller window. - Limits: BLS Occupational Outlook Handbook, Tellers, Wayback copy (2026-09-13). - BLS Monthly Labor Review 2006 - Wyatt and Hecker, Occupational changes during the 20th century (Wayback) / lines 1399–1400 - Selected supporting excerpt: database software, bookkeepers; ATM’s and telephone and online banking, tellers; - Limits: BLS Monthly Labor Review, March 2006 (Wyatt and Hecker), Wayback copy. Based on decennial Census data. ### h-farms / measured / reviewed Label: Official estimate In 1900, 41 of every 100 American workers farmed. By 2000, about 2 did. Scope: Share of the U.S. workforce in agriculture, 1900 and 2000 (USDA ERS, 2005) Limitations: 1900 counts the whole workforce (41%); 2000 counts employed workers (1.9%). On the page 1 figure = 1 in 100 workers: 41 then 2 in the wheat, 59 then 98 at the town edge. A century apart, these are not the same people. - USDA ERS - The 20th Century Transformation of U.S. Agriculture and Farm Policy (EIB-3, 2005) / lines 169–171 - Selected supporting excerpt: 1900 41 percent of workforce employed in agriculture - Limits: USDA Economic Research Service, Economic Information Bulletin 3 (June 2005), text extracted from the PDF. - USDA ERS - The 20th Century Transformation of U.S. Agriculture and Farm Policy (EIB-3, 2005) / lines 191–194 - Selected supporting excerpt: 2000/02 1.9 percent of employed labor force worked in agriculture (2000) - Limits: USDA Economic Research Service, Economic Information Bulletin 3 (June 2005), text extracted from the PDF. - USDA ERS - The 20th Century Transformation of U.S. Agriculture and Farm Policy (EIB-3, 2005) / lines 199–200 - Selected supporting excerpt: Share of workforce employed in agriculture, for 1900-1970, Historical - Limits: USDA Economic Research Service, Economic Information Bulletin 3 (June 2005), text extracted from the PDF. ### h-farm-output / measured / reviewed Label: Official estimate Farms now grow nearly three times what they did in 1948. Scope: Total U.S. farm output, 1948 to 2023 (USDA ERS) Limitations: Output of all farm production in 2023 compared with 1948. The taller wheat on the page is an illustration of that ratio, not crop height. - USDA ERS - Agricultural Productivity in the U.S., Summary of Recent Findings (updated 2026-08-25) / line 160 - Selected supporting excerpt: The level of U.S. farm output was nearly three times its 1948 level in 2023, growing at an average annual rate of 1.52 percent over the period. - Limits: USDA Economic Research Service web page, updated 2026-08-25, saved 2026-10-05. ### h-looms / documented / reviewed Label: Attribution Power looms took over 98% of the work needed to weave each yard of cloth. Weaving jobs still grew. Scope: 19th-century weaving (Bessen, IMF Finance & Development, March 2015) Limitations: Economist's summary, not a measurement we checked. - IMF F&D 2015 - Bessen, Toil and Technology (Wayback) / line 75 - Selected supporting excerpt: During the 19th century, for example, power looms automated 98 percent of the labor needed to weave a yard of cloth. Yet the number of factory weaving jobs increased over this period. - Limits: James Bessen's article in the IMF's Finance & Development (March 2015), Wayback copy. An economist's summary, not data we checked. - IMF F&D 2015 - Bessen, Toil and Technology (Wayback) / line 123 - Selected supporting excerpt: James Bessen is Lecturer in Law at the Boston University School of Law - Limits: James Bessen's article in the IMF's Finance & Development (March 2015), Wayback copy. An economist's summary, not data we checked. ### h-tool-stayed / editorial / reviewed Label: Editorial In every example here, the new tool stayed. What changed was the work. Scope: Our view of the history examples: computers, switchboards, tellers, farms and looms Limitations: Our view. It does not say every worker was better off, or that the same people got the new work. - IMF F&D 2015 - Bessen, Toil and Technology (Wayback) / line 77 - Selected supporting excerpt: In each case, the new jobs required new and different skills. - Limits: James Bessen's article in the IMF's Finance & Development (March 2015), Wayback copy. An economist's summary, not data we checked. ### h-asphalt-1890s / documented / reviewed Label: Official description Asphalt shingles were already in use in the 1890s. Scope: Roofing materials in the U.S. (National Park Service) Limitations: The NPS brief says asphalt shingles and roll roofing "were used in the 1890s". It does not say they were new then, so the tap-card line "were new in the 1890s" should read like this claim. - NPS - Preservation Brief 4, Roofing for Historic Buildings / lines 205–207 - Selected supporting excerpt: Other Materials: Asphalt shingles and roll roofing were used in the 1890s. - Limits: National Park Service Preservation Brief 4, text extracted from the PDF. ### h-roofers-history / measured / reviewed Label: Official estimate About 9,000 people worked as roofers in 1900 and 65,000 in 1970. Roofers held about 167,000 jobs in 2025. On the street, 1 figure = 1,000 roofers: 9, then 65, then 167. Scope: U.S. roofers: Census 1900 and 1970 (Historical Statistics, Series D439) and BLS 2025 Limitations: Census counted people, BLS counts jobs, and the country grew too. The Census series is 'Roofers and slaters' (1970 column: ages 16 and over). BLS: about 166,900 jobs in 2025, shown as 167 figures. Your figure is not part of the count. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1783 - Selected supporting excerpt: Series D 233-682. Detailed Occupation of the Economically Active Population: 1900 to 1970—Con. - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1784 - Selected supporting excerpt: [In thousands of persons 14 years old and over, except as indicated] - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1788 - Selected supporting excerpt: No. 16 years | 14 yearsj; 1970 1960 1960 1950 1940 1930 1920 1910 1900 - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. - Census - Historical Statistics of the US, Ch. D labor (OCR text) / line 1872 - Selected supporting excerpt: 439 Roofers and slaters._.__..___-..-----.---u------------ 65 65 5T 55; 49 50: 33 24 12 14 9 - Limits: OCR (tesseract) of the scanned 1975 Census volume; the OCR string is kept exactly, including misreads. Values were checked against the page image. Counts are thousands of people in the economically active population; the 1970 figure used is the column for people 16 and over. Row D439: 65 is the 1970 value (ages 16 and over) and 9 the 1900 value; OCR reads 57 as 5T. - U.S. BLS Occupational Outlook Handbook - Roofers (Wayback capture 2026-09-20) / line 268 - Selected supporting excerpt: Roofers held about 166,900 jobs in 2025. - Limits: BLS Occupational Outlook Handbook, Roofers, Wayback copy (2026-09-20); the live page could not be saved. ### h-fortified / self-reported / reviewed Label: Vendor-reported FORTIFIED, a stronger roof standard, passed 100,000 designations in 2026. More than half were earned in the last three years. Scope: IBHS FORTIFIED program, United States, to May 2026 Limitations: The program's own count (IBHS runs FORTIFIED). A designation is for one home or building; the release counts designations, so a page line saying "100,000 homes and buildings" paraphrases it. FORTIFIED is a voluntary, beyond-code construction and re-roofing method. - IBHS - FORTIFIED program surpasses 100,000 designations (news release, 2026-05-20) / line 222 - Selected supporting excerpt: RICHBURG, S.C., May 20, 2026 — The Insurance Institute for Business & Home Safety (IBHS) today announced its FORTIFIED® program has surpassed 100,000 designations nationwide - Limits: IBHS news release, saved 2026-10-05. IBHS runs the FORTIFIED program. - IBHS - FORTIFIED program surpasses 100,000 designations (news release, 2026-05-20) / line 226 - Selected supporting excerpt: FORTIFIED is a voluntary, beyond-code construction and re-roofing method that strengthens the parts of homes and buildings most vulnerable to high winds and wind‑driven rain. - Limits: IBHS news release, saved 2026-10-05. IBHS runs the FORTIFIED program. - IBHS - FORTIFIED program surpasses 100,000 designations (news release, 2026-05-20) / line 226 - Selected supporting excerpt: More than half of all FORTIFIED designations have been earned in just the last three years - Limits: IBHS news release, saved 2026-10-05. IBHS runs the FORTIFIED program. - IBHS - FORTIFIED program surpasses 100,000 designations (news release, 2026-05-20) / line 260 - Selected supporting excerpt: Originally focused on single‑family homes, the FORTIFIED program has expanded to include commercial buildings and multifamily communities - Limits: IBHS news release, saved 2026-10-05. IBHS runs the FORTIFIED program. ### h-fortified-sally / measured / reviewed Label: Official estimate After Hurricane Sally, FORTIFIED homes in Alabama were 55% to 74% less likely to have a loss. Scope: Insured homes in coastal Alabama after Hurricane Sally (2020) Limitations: A study by the Alabama Department of Insurance and the University of Alabama (May 2025) of more than 40,000 insured coastal homes. "Less likely to have a loss" is loss frequency (claims per policy) compared with conventional homes; the range depends on the designation level, the sample and the comparison. IBHS shared data and technical expertise with the authors. - Alabama Dept. of Insurance and University of Alabama CRIR - Performance of IBHS FORTIFIED Home construction in Hurricane Sally (2025) / lines 171–173 - Selected supporting excerpt: Alabama Department of Insurance and Center for Risk and Insurance Research – University of Alabama May 5, 2025 - Limits: Study report by the Alabama Department of Insurance and the University of Alabama Center for Risk and Insurance Research (May 5, 2025), text extracted from the PDF. - Alabama Dept. of Insurance and University of Alabama CRIR - Performance of IBHS FORTIFIED Home construction in Hurricane Sally (2025) / line 226 - Selected supporting excerpt: Claim frequency is the number of claims divided by the number of policies. - Limits: Study report by the Alabama Department of Insurance and the University of Alabama Center for Risk and Insurance Research (May 5, 2025), text extracted from the PDF. - Alabama Dept. of Insurance and University of Alabama CRIR - Performance of IBHS FORTIFIED Home construction in Hurricane Sally (2025) / line 235 - Selected supporting excerpt: reduced loss frequency by 55% to 74%, loss severity by 14% to 40%, and loss ratio by 51% to 72%. - Limits: Study report by the Alabama Department of Insurance and the University of Alabama Center for Risk and Insurance Research (May 5, 2025), text extracted from the PDF. - IBHS - Study shows FORTIFIED reduced Hurricane Sally damage (news release) / line 246 - Selected supporting excerpt: The study sample includes information from more than 40,000 insured properties in coastal Alabama. - Limits: IBHS news release, saved 2026-10-05. IBHS runs the FORTIFIED program. ### h-fortified-ibhs-role / self-reported / reviewed Label: Vendor-reported IBHS shared data with the Hurricane Sally study. Insurers have played a central role in backing FORTIFIED. Scope: Who is behind FORTIFIED and the Sally study Limitations: The saved sources do not say how IBHS itself is funded. The tap-card line "IBHS is funded by insurers" needs another saved source (for example IBHS's own About page) or should change to this wording. - Alabama Dept. of Insurance and University of Alabama CRIR - Performance of IBHS FORTIFIED Home construction in Hurricane Sally (2025) / lines 368–369 - Selected supporting excerpt: We especially appreciate the teams at IBHS, RenaissanceRe Risk Sciences, and Smart Home America for sharing data and technical expertise. - Limits: Study report by the Alabama Department of Insurance and the University of Alabama Center for Risk and Insurance Research (May 5, 2025), text extracted from the PDF. - IBHS - FORTIFIED program surpasses 100,000 designations (news release, 2026-05-20) / line 252 - Selected supporting excerpt: Insurers have played a central role in accelerating national adoption of FORTIFIED, supporting research, incentives, education, and program delivery - Limits: IBHS news release, saved 2026-10-05. IBHS runs the FORTIFIED program. ### a-nasa-1943 / documented / reviewed Label: Official description Archival photo · NACA / NASA Ames Research Center, ARC-1943-AAL-4961 (1943): human computers processing wind tunnel data. Scope: Caption and credit for the 1943 computing-room photo Limitations: From NASA's image record. NASA media are generally not subject to U.S. copyright; credit NASA and never imply endorsement. Shown untouched: camera motion only; never drawn on, animated or AI-processed. - NASA Image Library - ARC-1943-AAL-4961 record (API) / line 29 - Selected supporting excerpt: "title": "Computers processing data from the 16 Foot Wind Tunnel at Ames Research Center." - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - ARC-1943-AAL-4961 record (API) / line 14 - Selected supporting excerpt: "date_created": "1943-01-05T00:00:00Z" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - ARC-1943-AAL-4961 record (API) / line 28 - Selected supporting excerpt: "photographer": "NACA" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - ARC-1943-AAL-4961 record (API) / line 27 - Selected supporting excerpt: "nasa_id": "ARC-1943-AAL-4961" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 717 - Selected supporting excerpt: generally are not subject to copyright in the United States. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 719 - Selected supporting excerpt: NASA content used in a factual manner that does not imply endorsement may be used without needing explicit permission. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 735 - Selected supporting excerpt: For use of NASA images clearances may be necessary for images that include any NASA logos or NASA employees to be used as cover art or in promotional content. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. ### a-nasa-1952 / documented / reviewed Label: Official description Archival photo · NACA Langley / NASA, LRC-1952-B701_P-74769 (1952): a human computer with a Friden calculating machine. Scope: Caption and credit for the 1952 photo of a human computer Limitations: From NASA's image record. NASA media are generally not subject to U.S. copyright; credit NASA and never imply endorsement. Shown untouched: camera motion only; never drawn on, animated or AI-processed. NASA's description calls it a publicity photo from Muroc, California. - NASA Image Library - LRC-1952-B701_P-74769 record (API) / line 26 - Selected supporting excerpt: "title": "Langely Women Computer with X-4 with Friden machine" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - LRC-1952-B701_P-74769 record (API) / line 11 - Selected supporting excerpt: "date_created": "1952-03-24T00:00:00Z" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - LRC-1952-B701_P-74769 record (API) / line 25 - Selected supporting excerpt: "photographer": "NACA Langley" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - LRC-1952-B701_P-74769 record (API) / line 24 - Selected supporting excerpt: "nasa_id": "LRC-1952-B701_P-74769" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - LRC-1952-B701_P-74769 record (API) / line 12 - Selected supporting excerpt: Publicity photo from Muroc California, showing female support personnel with equipment. - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 717 - Selected supporting excerpt: generally are not subject to copyright in the United States. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 719 - Selected supporting excerpt: NASA content used in a factual manner that does not imply endorsement may be used without needing explicit permission. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 735 - Selected supporting excerpt: For use of NASA images clearances may be necessary for images that include any NASA logos or NASA employees to be used as cover art or in promotional content. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. ### a-nasa-1957 / documented / reviewed Label: Official description Archival photo · NACA Langley / NASA, photo by Taub, LRC-1957-B701_P-05391 (1957): IBM 704 computer operations. Scope: Caption and credit for the 1957 IBM 704 room photo Limitations: From NASA's image record. NASA media are generally not subject to U.S. copyright; credit NASA and never imply endorsement. Shown untouched: camera motion only; never drawn on, animated or AI-processed. NASA's record names the people in the photo; the page does not name them. - NASA Image Library - LRC-1957-B701_P-05391 record (API) / line 27 - Selected supporting excerpt: "title": "IBM 704 Computer Operations " - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - LRC-1957-B701_P-05391 record (API) / line 11 - Selected supporting excerpt: "date_created": "1957-12-16T00:00:00Z" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - LRC-1957-B701_P-05391 record (API) / line 26 - Selected supporting excerpt: "photographer": "NACA, Taub" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA Image Library - LRC-1957-B701_P-05391 record (API) / line 25 - Selected supporting excerpt: "nasa_id": "LRC-1957-B701_P-05391" - Limits: NASA Image and Video Library API record for this image, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 717 - Selected supporting excerpt: generally are not subject to copyright in the United States. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 719 - Selected supporting excerpt: NASA content used in a factual manner that does not imply endorsement may be used without needing explicit permission. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. - NASA - Images and media usage guidelines (Brand Center) / line 735 - Selected supporting excerpt: For use of NASA images clearances may be necessary for images that include any NASA logos or NASA employees to be used as cover art or in promotional content. - Limits: NASA's images and media usage guidelines page, saved 2026-10-05. ### a-loc-1959 / documented / reviewed Label: Official description Archival photo · Library of Congress, Prints & Photographs Division, U.S. News & World Report Collection, LC-DIG-ds-04925 (Marion S. Trikosko, 1959): women working at the U.S. Capitol switchboard. Scope: Caption and credit for the 1959 switchboard photo Limitations: From the Library of Congress record. Rights advisory: no known restrictions on publication. The Library does not grant permission; rights assessment is the user's. Camera motion only; never drawn on, animated or AI-processed. The title was written by Library staff. - Library of Congress - Item record, U.S. Capitol switchboard 1959, LC-DIG-ds-04925 (Wayback) / line 122 - Selected supporting excerpt: Title: [Women working at the U.S. Capitol switchboard, Washington, D.C.] - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, U.S. Capitol switchboard 1959, LC-DIG-ds-04925 (Wayback) / line 126 - Selected supporting excerpt: Trikosko, Marion S., photographer - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, U.S. Capitol switchboard 1959, LC-DIG-ds-04925 (Wayback) / line 129 - Selected supporting excerpt: 1959 Jan. 27. - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, U.S. Capitol switchboard 1959, LC-DIG-ds-04925 (Wayback) / line 137 - Selected supporting excerpt: LC-DIG-ds-04925 (digital file from original negative) - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, U.S. Capitol switchboard 1959, LC-DIG-ds-04925 (Wayback) / line 140 - Selected supporting excerpt: No known restrictions on publication. - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, U.S. Capitol switchboard 1959, LC-DIG-ds-04925 (Wayback) / line 154 - Selected supporting excerpt: U.S. News & World Report Magazine Photograph Collection. - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). ### a-loc-1908 / documented / reviewed Label: Official description Archival photo · Library of Congress, LC-USZ62-23640 (1908): bookkeeper's and teller's department, First National Bank of Somerville, New Jersey. Scope: Caption and credit for the 1908 bank photo Limitations: From the Library of Congress record. Rights advisory: no known restrictions on publication. The Library does not grant permission; rights assessment is the user's. Camera motion only; never drawn on, animated or AI-processed. Low-resolution copy negative: shown framed and small. - Library of Congress - Item record, Somerville bank 1908, LC-USZ62-23640 (Wayback) / line 113 - Selected supporting excerpt: Title: Bookkeeper's and teller's department First National Bank of Somerville, New Jersey - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, Somerville bank 1908, LC-USZ62-23640 (Wayback) / line 116 - Selected supporting excerpt: 1908. - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, Somerville bank 1908, LC-USZ62-23640 (Wayback) / line 127 - Selected supporting excerpt: LC-USZ62-23640 (b&w film copy neg.) - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, Somerville bank 1908, LC-USZ62-23640 (Wayback) / line 130 - Selected supporting excerpt: No known restrictions on publication. - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). ### a-loc-1937 / documented / reviewed Label: Official description Archival photo · Library of Congress, Prints & Photographs Division, Harris & Ewing Collection, LC-DIG-hec-22947 (1937): three operators at the Boy Scout Jamboree switchboard, Washington, D.C. Scope: Caption and credit for the 1937 switchboard photo Limitations: From the Library of Congress record. Rights advisory: no known restrictions on publication. The Library does not grant permission; rights assessment is the user's. Camera motion only; never drawn on, animated or AI-processed. Title from unverified caption data received with the collection. - Library of Congress - Item record, Jamboree switchboard 1937, LC-DIG-hec-22947 (Wayback) / line 132 - Selected supporting excerpt: Title: Boy Scout Jamboree switchboard. Washington D.C June 29. - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, Jamboree switchboard 1937, LC-DIG-hec-22947 (Wayback) / line 136 - Selected supporting excerpt: Harris & Ewing, photographer - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, Jamboree switchboard 1937, LC-DIG-hec-22947 (Wayback) / line 139 - Selected supporting excerpt: [19]37 June 29. - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, Jamboree switchboard 1937, LC-DIG-hec-22947 (Wayback) / line 147 - Selected supporting excerpt: LC-DIG-hec-22947 (digital file from original negative) - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). - Library of Congress - Item record, Jamboree switchboard 1937, LC-DIG-hec-22947 (Wayback) / line 150 - Selected supporting excerpt: No known restrictions on publication. For more information, see Harris & Ewing Photographs - Rights and Restrictions Information(http://www.loc.gov/rr/print/res/140_harr.html) - Limits: Library of Congress catalog record, Wayback copy (loc.gov itself was behind a bot check). ### m-office-pool / illustrative / reviewed Label: Model output In Foundation's model, about 4 of every 100 office workers are out of work or retraining at the start. By model year 6 it is about 4 in the Mild scenario, 5 in Moderate and 15 in Aggressive. Scope: Foundation's model: the office (knowledge-qualified) pool, start and model year 6 Limitations: Foundation's model, not a forecast. Model year 6 is a step in a made-up sequence, not a date. Builds on Korinek et al. (2026). Exact values: 3.8 at the start; 3.89, 4.59 and 15.10 per 100 at model year 6. 'Office workers' is the model's knowledge-qualified pool, and the rate includes people in training. Each figure is 1 in 100 of a made-up pool, not a real person. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 116–117 - Selected supporting excerpt: "scenario": "mild", "model_year": 0, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 123 - Selected supporting excerpt: "knowledge_jobseeker_unemployment_pct": 3.800000000000003, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 332–333 - Selected supporting excerpt: "scenario": "mild", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 339 - Selected supporting excerpt: "knowledge_jobseeker_unemployment_pct": 3.8924263707612345, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 584–585 - Selected supporting excerpt: "scenario": "moderate", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 591 - Selected supporting excerpt: "knowledge_jobseeker_unemployment_pct": 4.594208529854438, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 836–837 - Selected supporting excerpt: "scenario": "aggressive", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 843 - Selected supporting excerpt: "knowledge_jobseeker_unemployment_pct": 15.095952951735503, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - data README (README) / line 15 - Selected supporting excerpt: `knowledge_jobseeker_unemployment_pct` uses the current knowledge-qualified pool including training; - Limits: The paper's notes on its data files. - Foundation Projects - The work above us (article text, 2026-10-03) / line 326 - Selected supporting excerpt: Each group starts with 3.8% unemployment - Limits: The paper's article text, saved 2026-10-03. ### m-workforce / illustrative / reviewed Label: Model output Across all workers in Foundation's model, about 4 of every 100 are out of work or retraining at the start. By model year 6 it is about 4 in Mild, 4 in Moderate and 12 in Aggressive. Scope: Foundation's model: the whole synthetic workforce of 100, start and model year 6 Limitations: Foundation's model, not a forecast. Model year 6 is a step in a made-up sequence, not a date. Builds on Korinek et al. (2026). Exact values: 3.8 at the start; 3.79, 4.08 and 12.20 per 100 at model year 6. The rate includes people in training. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 116–117 - Selected supporting excerpt: "scenario": "mild", "model_year": 0, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 122 - Selected supporting excerpt: "unemployment_pct": 3.8000000000000043, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 332–333 - Selected supporting excerpt: "scenario": "mild", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 338 - Selected supporting excerpt: "unemployment_pct": 3.789132120464714, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 584–585 - Selected supporting excerpt: "scenario": "moderate", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 590 - Selected supporting excerpt: "unemployment_pct": 4.082272089087404, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 836–837 - Selected supporting excerpt: "scenario": "aggressive", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 842 - Selected supporting excerpt: "unemployment_pct": 12.200845135174452, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - data README (README) / line 15 - Selected supporting excerpt: The headline `unemployment_pct` already includes trainees, who remain active jobseekers in this model. - Limits: The paper's notes on its data files. ### m-roof-hours / illustrative / reviewed Label: Model output In Foundation's model at model year 6, machines do about 2 of every 100 hours of roof work in the Mild scenario, 14 in Moderate and 42 in Aggressive. People still put in about 98, 88 and 60 hours on the same roof. Scope: Foundation's model: one roof's work, model year 6 Limitations: Foundation's model, not a forecast. Model year 6 is a step in a made-up sequence, not a date. Builds on Korinek et al. (2026). Exact values: 2.0, 13.5 and 42.3 machine hours; human hours 0.98, 0.88 and 0.60 of the starting amount. Machine hours count human work replaced, not machine runtime. In the hardest scenario people still put in about 60 hours, more than the 57.7 machines leave, because the model adds back support hours. Hours, not people. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 332–333 - Selected supporting excerpt: "scenario": "mild", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 354 - Selected supporting excerpt: "human_hours_roof": 0.9766666666666667, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 355 - Selected supporting excerpt: "roof_robot_task_share_pct": 2.0, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 584–585 - Selected supporting excerpt: "scenario": "moderate", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 606 - Selected supporting excerpt: "human_hours_roof": 0.8843999999999999, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 607 - Selected supporting excerpt: "roof_robot_task_share_pct": 13.5, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 836–837 - Selected supporting excerpt: "scenario": "aggressive", "model_year": 6, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 858 - Selected supporting excerpt: "human_hours_roof": 0.5955722222222222, - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / line 859 - Selected supporting excerpt: "roof_robot_task_share_pct": 42.300000000000004, - Limits: The paper's published model output (scenario-data.json). Not a forecast. ### m-roof-zones / illustrative / reviewed Label: Scenario assumption One roof's 100 hours are split into the roofer's eight tasks by the authors' time weights: rig safe access 10, move materials 12, tear-off 14, deck repair 12, underlayment 12, install covering 22, flashing 12, inspect and hand over 6. Scope: One roof, 100 hours of work, in eight task zones Limitations: The authors' chosen weights for a mix of reroofing and repair, not a time study. Task names: see cl-onet-roofer. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 94–103 - Selected supporting excerpt: "roof_task_names": [ "access_safety", "materials", "tear_off", "deck_repair", "underlayment", "field_covering", "flashing", "handover" ], - Limits: The paper's published model output (scenario-data.json). Not a forecast. - Foundation Projects - The work above us - scenario data (scenario-data) / lines 104–113 - Selected supporting excerpt: "roof_task_weights": [ 0.1, 0.12, 0.14, 0.12, 0.12, 0.22, 0.12, 0.06 ], - Limits: The paper's published model output (scenario-data.json). Not a forecast. ### m-roof-zone-split / inferred / reviewed Label: Derived arithmetic In the hardest scenario the 42 machine hours fall by task: move materials 10, tear-off 6, deck repair 2, underlayment 6, install covering 16, flashing 1, inspect and hand over 1, rig safe access 0. Scope: Teal courses per zone on the roof Limitations: Our arithmetic: each task's weight x the authors' Aggressive machine share = 9.6, 6.3, 1.8, 6.0, 16.5, 1.2, 0.9 and 0 hours (42.3), rounded to whole courses by largest remainder so the zones add to 42. Moderate (14): move materials 4, tear-off 2, underlayment 2, install covering 6. Mild (2): move materials 1, install covering 1. Foundation's model, not a forecast. - Foundation Projects - The work above us - task bundles (task-bundles) / line 250 - Selected supporting excerpt: "baseline_time_weight": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 253 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 261 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 269 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 280 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 283 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 291 - Selected supporting excerpt: "automated_baseline_instances": 0.3, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 299 - Selected supporting excerpt: "automated_baseline_instances": 0.8, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 310 - Selected supporting excerpt: "baseline_time_weight": 0.14, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 313 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 321 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 329 - Selected supporting excerpt: "automated_baseline_instances": 0.45, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 340 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 343 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 351 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 359 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 370 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 373 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 381 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 389 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 400 - Selected supporting excerpt: "baseline_time_weight": 0.22, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 403 - Selected supporting excerpt: "automated_baseline_instances": 0.04, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 411 - Selected supporting excerpt: "automated_baseline_instances": 0.28, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 419 - Selected supporting excerpt: "automated_baseline_instances": 0.75, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 430 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 433 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 441 - Selected supporting excerpt: "automated_baseline_instances": 0.01, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 449 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 460 - Selected supporting excerpt: "baseline_time_weight": 0.06, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 463 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 471 - Selected supporting excerpt: "automated_baseline_instances": 0.03, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 479 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. ### m-roof-movers / inferred / reviewed Label: Derived arithmetic In the model, machines take the most hours from carrying materials and laying shingles. Rigging safe access stays with people in every scenario. Scope: Roofer's tasks, Foundation's model, model year 6 Limitations: Our arithmetic from the authors' shares: install covering and move materials have the most machine hours in all three scenarios; rig safe access is 0% in all three. Hours, not people. No task order: all roof tasks ramp together after model year 2. - Foundation Projects - The work above us - task bundles (task-bundles) / line 250 - Selected supporting excerpt: "baseline_time_weight": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 253 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 261 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 269 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 280 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 283 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 291 - Selected supporting excerpt: "automated_baseline_instances": 0.3, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 299 - Selected supporting excerpt: "automated_baseline_instances": 0.8, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 310 - Selected supporting excerpt: "baseline_time_weight": 0.14, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 313 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 321 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 329 - Selected supporting excerpt: "automated_baseline_instances": 0.45, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 340 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 343 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 351 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 359 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 370 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 373 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 381 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 389 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 400 - Selected supporting excerpt: "baseline_time_weight": 0.22, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 403 - Selected supporting excerpt: "automated_baseline_instances": 0.04, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 411 - Selected supporting excerpt: "automated_baseline_instances": 0.28, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 419 - Selected supporting excerpt: "automated_baseline_instances": 0.75, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 430 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 433 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 441 - Selected supporting excerpt: "automated_baseline_instances": 0.01, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 449 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 460 - Selected supporting excerpt: "baseline_time_weight": 0.06, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 463 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 471 - Selected supporting excerpt: "automated_baseline_instances": 0.03, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 479 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. ### m-pm-agent-share / inferred / reviewed Label: Derived arithmetic Take the production manager's 100 office hours: AI agents do about 11 in the Mild scenario, 32 in Moderate and 55 in Aggressive. Scope: The production manager's 100 hours on the office wall, model year 6 Limitations: Our arithmetic: each task's time weight x the authors' share done by AI agents, summed: 11.45, 31.82 and 55.25 hours per 100. The shares are the authors' assumptions. Hours, not people. Foundation's model, not a forecast. - Foundation Projects - The work above us - task bundles (task-bundles) / line 8 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 11 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 19 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 27 - Selected supporting excerpt: "automated_baseline_instances": 0.35, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 38 - Selected supporting excerpt: "baseline_time_weight": 0.16, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 41 - Selected supporting excerpt: "automated_baseline_instances": 0.2, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 49 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 57 - Selected supporting excerpt: "automated_baseline_instances": 0.8, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 68 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 71 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 79 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 87 - Selected supporting excerpt: "automated_baseline_instances": 0.85, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 98 - Selected supporting excerpt: "baseline_time_weight": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 101 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 109 - Selected supporting excerpt: "automated_baseline_instances": 0.4, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 117 - Selected supporting excerpt: "automated_baseline_instances": 0.75, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 128 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 131 - Selected supporting excerpt: "automated_baseline_instances": 0.25, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 139 - Selected supporting excerpt: "automated_baseline_instances": 0.7, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 147 - Selected supporting excerpt: "automated_baseline_instances": 0.95, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 158 - Selected supporting excerpt: "baseline_time_weight": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 161 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 169 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 177 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 188 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 191 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 199 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 207 - Selected supporting excerpt: "automated_baseline_instances": 0.3, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 218 - Selected supporting excerpt: "baseline_time_weight": 0.06, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 221 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 229 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 237 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. ### m-pm-new-hours / inferred / reviewed Label: Derived arithmetic The model adds new human work for the production manager: about 6 hours per 100 in Mild, 8 in Moderate and none in Aggressive. Scope: Gold cards on the office wall Limitations: Our arithmetic: weight x AI share x new human hours per automated hour (0.5, 0.25, 0), summed: 5.725, 7.955 and 0, shown as 6, 8 and 0 gold cards. Gold is hours of new work, not new hires. Foundation's model, not a forecast. - Foundation Projects - The work above us - task bundles (task-bundles) / line 8 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 16 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.5 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 24 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.25 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 32 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 38 - Selected supporting excerpt: "baseline_time_weight": 0.16, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 46 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.5 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 54 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.25 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 62 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 68 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 76 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.5 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 84 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.25 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 92 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 98 - Selected supporting excerpt: "baseline_time_weight": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 106 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.5 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 114 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.25 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 122 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 128 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 136 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.5 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 144 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.25 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 152 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 158 - Selected supporting excerpt: "baseline_time_weight": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 166 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.5 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 174 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.25 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 182 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 188 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 196 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.5 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 204 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.25 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 212 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 218 - Selected supporting excerpt: "baseline_time_weight": 0.06, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 226 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.5 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 234 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0.25 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 242 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 11 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 19 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 27 - Selected supporting excerpt: "automated_baseline_instances": 0.35, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 41 - Selected supporting excerpt: "automated_baseline_instances": 0.2, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 49 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 57 - Selected supporting excerpt: "automated_baseline_instances": 0.8, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 71 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 79 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 87 - Selected supporting excerpt: "automated_baseline_instances": 0.85, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 101 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 109 - Selected supporting excerpt: "automated_baseline_instances": 0.4, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 117 - Selected supporting excerpt: "automated_baseline_instances": 0.75, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 131 - Selected supporting excerpt: "automated_baseline_instances": 0.25, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 139 - Selected supporting excerpt: "automated_baseline_instances": 0.7, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 147 - Selected supporting excerpt: "automated_baseline_instances": 0.95, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 161 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 169 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 177 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 191 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 199 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 207 - Selected supporting excerpt: "automated_baseline_instances": 0.3, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 221 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 229 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 237 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - data README (README) / line 21 - Selected supporting excerpt: New work is zero in the Aggressive case; the model does not guarantee new occupations or hires. - Limits: The paper's notes on its data files. ### m-office-movers / inferred / reviewed Label: Derived arithmetic In our model, records, estimates, orders and schedules move the most. Safety checks and surprises stay mostly with people. Scope: Production manager tasks, model year 6 Limitations: Our arithmetic from the authors' shares: track costs and progress, build the estimate, order materials and schedule the crew have the most AI-agent hours in every scenario (Aggressive: 11.4, 12.8, 10.2 and 11.25 of 100). Check safety and quality (10%) and handle surprises (5%) stay mostly with people even in the hardest scenario. Foundation's model, not a forecast. - Foundation Projects - The work above us - task bundles (task-bundles) / line 8 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 11 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 19 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 27 - Selected supporting excerpt: "automated_baseline_instances": 0.35, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 38 - Selected supporting excerpt: "baseline_time_weight": 0.16, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 41 - Selected supporting excerpt: "automated_baseline_instances": 0.2, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 49 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 57 - Selected supporting excerpt: "automated_baseline_instances": 0.8, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 68 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 71 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 79 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 87 - Selected supporting excerpt: "automated_baseline_instances": 0.85, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 98 - Selected supporting excerpt: "baseline_time_weight": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 101 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 109 - Selected supporting excerpt: "automated_baseline_instances": 0.4, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 117 - Selected supporting excerpt: "automated_baseline_instances": 0.75, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 128 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 131 - Selected supporting excerpt: "automated_baseline_instances": 0.25, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 139 - Selected supporting excerpt: "automated_baseline_instances": 0.7, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 147 - Selected supporting excerpt: "automated_baseline_instances": 0.95, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 158 - Selected supporting excerpt: "baseline_time_weight": 0.15, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 161 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 169 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 177 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 188 - Selected supporting excerpt: "baseline_time_weight": 0.12, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 191 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 199 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 207 - Selected supporting excerpt: "automated_baseline_instances": 0.3, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 218 - Selected supporting excerpt: "baseline_time_weight": 0.06, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 221 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 229 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. - Foundation Projects - The work above us - task bundles (task-bundles) / line 237 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task bundles: the authors' chosen time weights and shares. Not measured. ### m-season-table / inferred / reviewed Label: Derived arithmetic + Authored example Example seasons of 6, 8, 10 and 12 months are about 960, 1,280, 1,600 and 1,920 working hours. Against a 2,600-hour factory year, a machine's fixed cost per working hour is about 2.7, 2.0, 1.6 and 1.35 times as high. Scope: Season arithmetic for owners Limitations: Our arithmetic: months x 20 days x 8 hours; 2,600 ÷ hours (2.71, 2.03, 1.63, 1.35). Example seasons, not data. Assumes the machine costs the same per year either way (see cl-fixed-cost-ratio, cl-factory-hours). - Figure AI - Production at BMW (figure.ai, fetched 2026-10-03) / line 13 - Selected supporting excerpt: Ran 10-hour shift Monday-Friday - Limits: Figure's own report on its BMW deployment. ### fp-take-part / editorial / reviewed Label: Foundation position Foundation gives roofing owners, office teams and crews a way to take part. Scope: Closing line Limitations: A Foundation position, not evidence. ### fp-map-my-exit / editorial / reviewed Label: Project decision Map My Exit: free 30-minute call for roofing owners. Scope: Call to action for roofing owners Limitations: Foundation's own description of its offer, not evidence of any result. ### v3-view-storm / editorial / reviewed Label: Editorial Change is coming to roofing. You can't stop it. Scope: Opening line Limitations: Our view, not a finding and not the model. The paper calls the overlap of AI and robots in roofing 'possible, not inevitable'; this line is ours, not the paper's. - Foundation Projects - The work above us (article text, 2026-10-03) / line 269 - Selected supporting excerpt: That overlap is possible, not inevitable. - Limits: The paper's article text, saved 2026-10-03. ### v3-view-take-part / editorial / reviewed Label: Editorial You can take part in it. Or you can let it happen to you. Scope: Role pick Limitations: Our view, not a finding and not the model. ### v3-view-choice / editorial / reviewed Label: Editorial Nobody can stop this. Not us, not anyone. You decide what you do next. Taking part means learning the new tools early, and having a say in how they're used. Waiting is a choice too. The change comes anyway. The model's 15 don't change. What you do next does. Scope: The choice Limitations: Our view, not a finding and not the model. The gesture changes no model number. ### v3-view-trust / editorial / reviewed Label: Editorial AI can draft the estimate. People still earn the customer's trust. Scope: Office path Limitations: Our view, not a finding and not the model. ### v3-view-shape / editorial / reviewed Label: Editorial Roofs, materials and tools will change. The people who take part get to shape how. Scope: Imagined future Limitations: Our view, not a finding and not the model. ### v3-view-bigger / editorial / reviewed Label: Editorial Better tools let you take on bigger jobs. More ambition. A bigger vision. Scope: Imagined future: bigger jobs Limitations: Our view, not a finding and not the model. ### v3-view-close / editorial / reviewed Label: Editorial Don't put your head in the sand. Take part. Scope: Closing line Limitations: Our view, not a finding and not the model. ### v3-view-plan / editorial / reviewed Label: Editorial My plan to take part. Owner: let my crew help pick the first task to test; ask every robot vendor for cost per accepted task. Office team: learn to run the AI tools and check their work; own the customer conversation. Crew: learn the machines before they show up; teach the next roofer; ask who owns the recordings of my work. Scope: Plan card lines Limitations: Our view, not a finding and not the model. The card carries no model number. ### cl-frame / editorial / reviewed Label: Editorial framing AI is changing office work. Robots may change roof work. That is the paper's framing. Scope: The paper's opening question Limitations: Framing, not a finding. - Foundation Projects - The work above us (article text, 2026-10-03) / line 9 - Selected supporting excerpt: AI is changing the office. Robotics may change the roof. What happens when both transitions arrive together? - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-protection-period / editorial / reviewed Label: Editorial AI and robots could reach roofing at the same time, but that is not certain. The question is whether roof work stays the same long enough for people to learn the trade. Scope: The question behind the chart Limitations: A question, not a finding. - Foundation Projects - The work above us (article text, 2026-10-03) / line 269 - Selected supporting excerpt: That overlap is possible, not inevitable. We do not have an established date for widespread autonomous roofing. We do have reason to examine what would happen if the period of protection offered by physical work proved shorter than the time people needed to retrain. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-rufus / self-reported / reviewed Label: Partner announcement + Vendor-reported Saint-Gobain, the parent company of CertainTeed, announced a partnership with Renovate Robotics. Renovate's first robot installs asphalt shingles on steep roofs, and pilot projects were planned with contractors. Renovate calls its robot, Rufus, the world's first shingle installation robot. Scope: Shingle-installing robots Limitations: Does not show finished pilots, a whole roof done on its own, or installed cost. 'World's first' is the company's own claim. The live announcement page could not be saved; this is a Wayback copy. - Saint-Gobain North America - Partnership with Renovate Robotics (Wayback capture 2025-07-16) / line 23 - Selected supporting excerpt: Saint-Gobain, the parent company of CertainTeed - Limits: Saint-Gobain's announcement, Wayback copy; the live page could not be saved. - Saint-Gobain North America - Partnership with Renovate Robotics (Wayback capture 2025-07-16) / line 24 - Selected supporting excerpt: Renovate’s first robot automates asphalt shingle installation on steep slope roofs. Under the partnership, Renovate will work with roofing professionals from within CertainTeed’s network of credentialed contractors to complete pilot projects with its robot. - Limits: Saint-Gobain's announcement, Wayback copy; the live page could not be saved. - Renovate Robotics - Home page (Wayback capture 2026-09-18) / line 27 - Selected supporting excerpt: We built Rufus, the world’s first shingle installation robot - Limits: Renovate's own home page, Wayback copy. - Foundation Projects - The work above us (article text, 2026-10-03) / line 283 - Selected supporting excerpt: Saint-Gobain announced a partnership with Renovate Robotics - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-rufus-demo / editorial / reviewed Label: Project decision Robots built only to install shingles have been announced and demonstrated. No robot has been shown doing a whole roof on its own. Any cost for a robot doing a roof task is a model, not a real-world measurement. Scope: How this page describes roof robots Limitations: The demonstration was a prototype run by a technician with a tablet, on a model roof in a warehouse. It was not a customer roof and not a robot working on its own. - Qualified Remodeler - Renovate Robotics reveals roofing robot prototype (trade press, fetched 2026-10-04) / line 72 - Selected supporting excerpt: The video shows a robot, operated by a technician via a tablet, installing asphalt shingles on a model roof in a warehouse environment. - Limits: A trade-press report on a company video. Not an independent test. ### cl-ramp-rule / illustrative / reviewed Label: Scenario assumption In the model, AI changes office work from the start. Machines start on roof tasks only after model year 2. Scope: Timing in the model Limitations: The authors' timing choice. Model years are not dates. - Foundation Projects - The work above us (article text, 2026-10-03) / line 327 - Selected supporting excerpt: Digital progress rises linearly over six periods. Direct robotics begins after period 2 and rises over four periods. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-task-shares / illustrative / reviewed Label: Scenario assumption The share of each task done by machines or AI agents at model year 6, in each scenario. The authors chose these shares. Scope: Task shares in the task grids and the 16-task lists Limitations: Assumptions for model year 6, not adoption estimates or measurements. Counted in hours of human work replaced, not machine running time. - Foundation Projects - The work above us - task bundles (task-bundles) / line 2 - Selected supporting excerpt: "classification": "author-assigned time weights and terminal deployment shares; not measured O*NET times", - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 11 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 19 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 27 - Selected supporting excerpt: "automated_baseline_instances": 0.35, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 41 - Selected supporting excerpt: "automated_baseline_instances": 0.2, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 49 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 57 - Selected supporting excerpt: "automated_baseline_instances": 0.8, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 71 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 79 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 87 - Selected supporting excerpt: "automated_baseline_instances": 0.85, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 101 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 109 - Selected supporting excerpt: "automated_baseline_instances": 0.4, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 117 - Selected supporting excerpt: "automated_baseline_instances": 0.75, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 131 - Selected supporting excerpt: "automated_baseline_instances": 0.25, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 139 - Selected supporting excerpt: "automated_baseline_instances": 0.7, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 147 - Selected supporting excerpt: "automated_baseline_instances": 0.95, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 161 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 169 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 177 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 191 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 199 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 207 - Selected supporting excerpt: "automated_baseline_instances": 0.3, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 221 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 229 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 237 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 253 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 261 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 269 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 283 - Selected supporting excerpt: "automated_baseline_instances": 0.05, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 291 - Selected supporting excerpt: "automated_baseline_instances": 0.3, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 299 - Selected supporting excerpt: "automated_baseline_instances": 0.8, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 313 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 321 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 329 - Selected supporting excerpt: "automated_baseline_instances": 0.45, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 343 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 351 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 359 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 373 - Selected supporting excerpt: "automated_baseline_instances": 0.02, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 381 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 389 - Selected supporting excerpt: "automated_baseline_instances": 0.5, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 403 - Selected supporting excerpt: "automated_baseline_instances": 0.04, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 411 - Selected supporting excerpt: "automated_baseline_instances": 0.28, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 419 - Selected supporting excerpt: "automated_baseline_instances": 0.75, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 433 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 441 - Selected supporting excerpt: "automated_baseline_instances": 0.01, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 449 - Selected supporting excerpt: "automated_baseline_instances": 0.1, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 463 - Selected supporting excerpt: "automated_baseline_instances": 0, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 471 - Selected supporting excerpt: "automated_baseline_instances": 0.03, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. - Foundation Projects - The work above us - task bundles (task-bundles) / line 479 - Selected supporting excerpt: "automated_baseline_instances": 0.15, - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. ### cl-anthropic-switch / documented / reviewed Label: Published model text Anthropic's Economics team built a model of how AI might affect jobs, growth and unemployment. In its bigger-change scenarios, coders and call-center agents may have to switch to jobs like electrician and nurse. Anthropic says switching jobs is difficult. Scope: Anthropic's bigger-change scenarios Limitations: Applies to their substantial and extreme scenarios only. Anthropic does not say switching jobs is easy. - Anthropic Institute - Economic scenarios for transformative AI (web page, fetched 2026-10-03) / line 199 - Selected supporting excerpt: In our substantial and extreme scenarios, knowledge workers may see a lot of automation and displacement. At the individual level, it means coders and call service center agents may have to switch to jobs like electrician and nurse, which are less exposed to AI. - Limits: Anthropic's public scenarios page, saved copy. - Anthropic Institute - Economic scenarios for transformative AI (web page, fetched 2026-10-03) / line 231 - Selected supporting excerpt: Switching to a new occupation is difficult for a few reasons - Limits: Anthropic's public scenarios page, saved copy. - Anthropic Institute - Economic scenarios for transformative AI (web page, fetched 2026-10-03) / line 14 - Selected supporting excerpt: Anthropic’s Economics team built a model of how AI might affect jobs, growth, and unemployment in the US in coming years. - Limits: Anthropic's public scenarios page, saved copy. ### cl-korinek-authors / documented / reviewed Label: Attribution This work builds on Economic Scenarios for Transformative AI by Anton Korinek, Charles I. Jones, Szymon Sacher, Tess Cotter and Peter McCrory (The Anthropic Institute Working Paper No. 2026-02). Scope: Credit Limitations: Their Modest, Substantial and Extreme scenarios belong to their model. The Mild, Moderate and Aggressive scenarios belong to Foundation's model, a different and simpler experiment. - Korinek, Jones, Sacher, Cotter, McCrory (2026) - Economic Scenarios for Transformative AI, Anthropic Institute WP 2026-02 (text) / line 1 - Selected supporting excerpt: Economic Scenarios for Transformative AI - Limits: The authors' working paper, saved text. - Korinek, Jones, Sacher, Cotter, McCrory (2026) - Economic Scenarios for Transformative AI, Anthropic Institute WP 2026-02 (text) / line 5 - Selected supporting excerpt: Anton Korinek, Charles I. Jones, Szymon Sacher, Tess Cotter, and Peter McCrory - Limits: The authors' working paper, saved text. - Korinek, Jones, Sacher, Cotter, McCrory (2026) - Economic Scenarios for Transformative AI, Anthropic Institute WP 2026-02 (text) / line 7 - Selected supporting excerpt: The Anthropic Institute - Limits: The authors' working paper, saved text. - Korinek, Jones, Sacher, Cotter, McCrory (2026) - Economic Scenarios for Transformative AI, Anthropic Institute WP 2026-02 (text) / line 8 - Selected supporting excerpt: Working Paper No. 2026-02 - Limits: The authors' working paper, saved text. - Anthropic Institute - Economic scenarios for transformative AI (web page, fetched 2026-10-03) / line 388 - Selected supporting excerpt: Anton Korinek, Chad Jones, Szymon Sacher, Tess Cotter, and Peter McCrory developed the economic model and co-authored the companion technical report. - Limits: Anthropic's public scenarios page, saved copy. ### cl-robotics-boundary / documented / reviewed Label: Published model text Anthropic's scenarios and the Korinek et al. working paper leave fast progress in robots out on purpose. Scope: What Anthropic's model covers Limitations: A stated model scope. It is not a claim that robots won't improve, and not a flaw. - Anthropic Institute - Economic scenarios for transformative AI (web page, fetched 2026-10-03) / line 381 - Selected supporting excerpt: For example, we did not include scenarios where humanity develops hyper-capable robots. - Limits: Anthropic's public scenarios page, saved copy. - Korinek, Jones, Sacher, Cotter, McCrory (2026) - Economic Scenarios for Transformative AI, Anthropic Institute WP 2026-02 (text) / line 209 - Selected supporting excerpt: We also consider only the effects of AI on cognitive tasks, and do not allow for rapid advances - Limits: The authors' working paper, saved text. - Korinek, Jones, Sacher, Cotter, McCrory (2026) - Economic Scenarios for Transformative AI, Anthropic Institute WP 2026-02 (text) / line 210 - Selected supporting excerpt: in robotics and their effects on physical tasks in the coming years. - Limits: The authors' working paper, saved text. ### cl-not-forecast / illustrative / reviewed Label: Scenario assumption The scenarios are examples, not forecasts. They have no calendar dates and no odds. Scope: All three scenarios Limitations: Model years are steps in the model, not calendar years. - Foundation Projects - The work above us (article text, 2026-10-03) / line 121 - Selected supporting excerpt: Illustrative scenarios, not forecasts. Model years have no assigned calendar dates. No scenario is assigned a probability. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-trades-time / editorial / reviewed Label: Editorial A move into the trades takes time. The work may change while a person is learning it. Scope: Moving into the trades Limitations: The paper's view. - Foundation Projects - The work above us (article text, 2026-10-03) / line 112 - Selected supporting excerpt: A move into the trades takes time. The work at the other end may change while a person is learning it. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-unitree / self-reported / reviewed Label: Vendor-reported Unitree says it shipped more than 5,500 human-shaped robots in 2025, sold and delivered. Shipping a robot is not the same as the robot doing useful work. Scope: Robot sales Limitations: The company's own number. - Unitree - Clarification regarding 2025 sales data (fetched 2026-10-03) / line 558 - Selected supporting excerpt: In 2025, Unitree's actual shipment volume of humanoid robots exceeded 5,500 units (referring to the quantity actually sold and delivered to end customers, not order volume; the order volume is higher). - Limits: Unitree's own statement. ### cl-bd-hand / self-reported / reviewed Label: Vendor-reported Boston Dynamics says its new Atlas robot hand can grip tools with triggers, such as nail guns. Scope: Robot hands Limitations: A hand that can grip a tool. Not accepted nailing or a watertight roof. - Boston Dynamics - Robot hands for modern AI and real work (2026-10-01, fetched 2026-10-03) / line 131 - Selected supporting excerpt: Triggered tool grasps, like drills, power torque drivers, grinders, nail guns, and welding torches - Limits: Boston Dynamics' own description of its product. - Foundation Projects - The work above us (article text, 2026-10-03) / line 284 - Selected supporting excerpt: a hand capable of gripping a triggered tool - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-bmw / self-reported / reviewed Label: Vendor-reported Figure says its robot at a BMW plant ran a 10-hour shift, Monday to Friday, loaded more than 90,000 parts and ran for more than 1,250 hours. Scope: A car factory, not a roof Limitations: Factory loading, not roof work. The company's own report; not checked by others. - Figure AI - Production at BMW (figure.ai, fetched 2026-10-03) / line 13 - Selected supporting excerpt: Ran 10-hour shift Monday-Friday - Limits: Figure's own page. - Figure AI - Production at BMW (figure.ai, fetched 2026-10-03) / line 14 - Selected supporting excerpt: 90,000+ parts loaded - Limits: Figure's own page. - Figure AI - Production at BMW (figure.ai, fetched 2026-10-03) / line 15 - Selected supporting excerpt: 1,250+ hours of runtime - Limits: Figure's own page. ### cl-onet-pm / documented / reviewed Label: Official description The production manager's eight task names come from O*NET task statements for construction managers (11-9021) and construction supervisors (47-1011). Scope: Names of the production manager's eight tasks Limitations: The task names are the paper's short versions of O*NET statements, in this order: Inspect the roof, Build the estimate, Order materials, Schedule the crew, Track costs & progress, Check safety & quality, Coordinate people, Handle surprises. O*NET does not set weekly hours or automation shares. - O-NET OnLine - 11-9021.00 Construction Managers (fetched 2026-10-03) / line 115 - Selected supporting excerpt: Direct how drone technology is used for site inspections and progress monitoring, ensuring accurate and timely project completion. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 47-1011.00 First-Line Supervisors of Construction Trades (fetched 2026-10-03) / line 96 - Selected supporting excerpt: Estimate material or worker requirements to complete jobs. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 47-1011.00 First-Line Supervisors of Construction Trades (fetched 2026-10-03) / line 98 - Selected supporting excerpt: Order or requisition materials or supplies. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 11-9021.00 Construction Managers (fetched 2026-10-03) / line 92 - Selected supporting excerpt: Plan, schedule, or coordinate construction project activities to meet deadlines. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 11-9021.00 Construction Managers (fetched 2026-10-03) / line 93 - Selected supporting excerpt: Prepare and submit budget estimates, progress reports, or cost tracking reports. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 47-1011.00 First-Line Supervisors of Construction Trades (fetched 2026-10-03) / line 91 - Selected supporting excerpt: Inspect work progress, equipment, or construction sites to verify safety or to ensure that specifications are met. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 47-1011.00 First-Line Supervisors of Construction Trades (fetched 2026-10-03) / line 93 - Selected supporting excerpt: Supervise, coordinate, or schedule the activities of construction or extractive workers. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 11-9021.00 Construction Managers (fetched 2026-10-03) / line 103 - Selected supporting excerpt: Implement new or modified plans in response to delays, bad weather, or construction site emergencies. - Limits: O*NET task statements. They set no weekly hours or automation shares. - Foundation Projects - The work above us (article text, 2026-10-03) / line 385 - Selected supporting excerpt: The descriptions do not establish our workweek weights or automation percentages. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-onet-roofer / documented / reviewed Label: Official description Five of the roofer's eight task names come from O*NET task statements for roofers (47-2181). Three are the authors' own task names. Scope: Names of the roofer's eight tasks Limitations: Set up safe access, Lay underlayment, Install roof covering, Fit the flashing and Inspect & hand over match O*NET statements (two are the closest match). Move materials, Remove old roofing and Repair the deck are the authors' own names. O*NET does not set weekly hours or automation shares. - O-NET OnLine - 47-2181.00 Roofers (fetched 2026-10-03) / line 100 - Selected supporting excerpt: Set up scaffolding to provide safe access to roofs. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 47-2181.00 Roofers (fetched 2026-10-03) / line 99 - Selected supporting excerpt: Attach roofing paper to roofs in overlapping strips to form bases for other materials. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 47-2181.00 Roofers (fetched 2026-10-03) / line 97 - Selected supporting excerpt: Cover roofs or exterior walls of structures with slate, asphalt, aluminum, wood, gravel, gypsum, or related materials, using brushes, knives, punches, hammers, or other tools. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 47-2181.00 Roofers (fetched 2026-10-03) / line 101 - Selected supporting excerpt: Cement or nail flashing strips of metal or shingle over joints to make them watertight. - Limits: O*NET task statements. They set no weekly hours or automation shares. - O-NET OnLine - 47-2181.00 Roofers (fetched 2026-10-03) / line 90 - Selected supporting excerpt: Inspect problem roofs to determine the best repair procedures. - Limits: O*NET task statements. They set no weekly hours or automation shares. ### cl-roof-no-autonomy / editorial / reviewed Label: Editorial No robot has been shown doing a whole roof on its own. There is no set date for robots doing roof work on their own. Scope: Roof robots today Limitations: The paper's reading of its sources. - Foundation Projects - The work above us (article text, 2026-10-03) / line 74 - Selected supporting excerpt: A scenario assumption. Today’s demonstrations do not establish whole-roof autonomy. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us (article text, 2026-10-03) / line 269 - Selected supporting excerpt: We do not have an established date for widespread autonomous roofing. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-evidence-boundaries / editorial / reviewed Label: Editorial What the robot reports do not show: a whole roof done on its own, reliability checked by others across many sites, or the full cost of an installed roof. Scope: All robot reports on this page Limitations: The paper's reading of its sources. - Foundation Projects - The work above us (article text, 2026-10-03) / line 283 - Selected supporting excerpt: These sources do not establish whole-roof autonomy, independently verified reliability across varied sites, or a complete installed-square cost. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us (article text, 2026-10-03) / line 284 - Selected supporting excerpt: is a component capability, not an accepted roof installation. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us (article text, 2026-10-03) / line 425 - Selected supporting excerpt: Holding a tool does not establish safe movement, accepted fastening, or a weatherproof roof. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-unemployment / illustrative / reviewed Label: Model output In the Aggressive scenario, unemployment in the national model reaches 12.2% at model year 6. Scope: Foundation's national model, Aggressive scenario Limitations: The count includes people in training. - Foundation Projects - The work above us (article text, 2026-10-03) / line 342 - Selected supporting excerpt: | National-model unemployment | % | 3.8 | 4.1 | 12.2 - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us - scenario data (scenario-data) / line 842 - Selected supporting excerpt: "unemployment_pct": 12.200845135174452, - Limits: The paper's published model output. Not a forecast. - Foundation Projects - The work above us - data README (README) / line 15 - Selected supporting excerpt: The headline `unemployment_pct` already includes trainees, who remain active jobseekers in this model. - Limits: The paper's notes on its data files. ### cl-machine-hours / illustrative / reviewed Label: Scenario assumption Hours done by machines or agents count the human work they replace, not how long the machine runs. Scope: What machine and agent hours mean Limitations: The paper's own definition. - Foundation Projects - The work above us (article text, 2026-10-03) / line 211 - Selected supporting excerpt: Digitally or robotically executed hours measure the original human workload replaced; they do not measure machine runtime. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-support-coeff / illustrative / reviewed Label: Scenario assumption For every hour of work machines take over, the model adds support time: 0.10 hours for office work and 0.15 hours for roof work. Scope: Support work in the task grids Limitations: Per hour of human work replaced, not machine running time. - Foundation Projects - The work above us (article text, 2026-10-03) / line 325 - Selected supporting excerpt: Support is assumed to be 0.10 per automated baseline hour for digital work and 0.15 for physical work. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-entrants / illustrative / reviewed Label: Scenario assumption By model year 6, new qualified people looking for roofing work equal 3%, 10% or 20% of starting roofing employment (Mild, Moderate, Aggressive). This is an assumption, not an observed flow. Scope: The new-workers switch Limitations: No headcount of new workers is shown. - Foundation Projects - The work above us (article text, 2026-10-03) / line 334 - Selected supporting excerpt: then adds qualified entrants of 3%, 10%, or 20% of baseline employment by year 6. Its 3.8% initial jobseeker share is an assumption, not a measured roofing rate. Entrants are not an empirically identified flow from the national model. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-wage-table / illustrative / reviewed Label: Model output At model year 6, office pay ends at 100.0, 99.6 and 95.3, and roofing pay at 98.5, 94.5 and 85.0 (Mild, Moderate, Aggressive). Scope: Pay at model year 6 Limitations: Model output from a chosen pay rule. - Foundation Projects - The work above us (article text, 2026-10-03) / line 339 - Selected supporting excerpt: | Knowledge-work wage | Index | 100.0 | 99.6 | 95.3 - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us (article text, 2026-10-03) / line 341 - Selected supporting excerpt: | Separate roofing wage | Index | 98.5 | 94.5 | 85.0 - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-wage-own-baseline / editorial / reviewed Label: Editorial Each pay index starts at 100 in its own job market. The numbers are not dollar comparisons. Scope: How to read the pay numbers Limitations: Office pay covers knowledge work. Roofing pay comes from a separate roofing model. - Foundation Projects - The work above us (article text, 2026-10-03) / line 143 - Selected supporting excerpt: ### Wages do not follow one path - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us (article text, 2026-10-03) / line 144 - Selected supporting excerpt: Each wage index starts at 100 in its own labor market; these are not dollar wage comparisons. Non-knowledge work is broader than trades; roofing is a separate module. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us - data README (README) / line 13 - Selected supporting excerpt: Each wage index is normalized to its own baseline of 100, so their levels do not compare absolute wage amounts. - Limits: The paper's notes on its data files. ### cl-wage-rule / illustrative / reviewed Label: Scenario assumption The way roofing pay responds in the model is a chosen rule, not an estimate from real roofing markets. Scope: The pay rule Limitations: The paper's own description of its model. - Foundation Projects - The work above us (article text, 2026-10-03) / line 290 - Selected supporting excerpt: Its wage response is a chosen rule, not an estimate from observed roofing markets. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-new-duties-text / editorial / reviewed Label: Editorial New work for people: setting up machines, supervising new equipment, and trying out safer methods, then teaching them. Scope: New duties Limitations: The paper's view. - Foundation Projects - The work above us (article text, 2026-10-03) / line 76 - Selected supporting excerpt: Setting up machines. Supervising unfamiliar equipment. Trialing a safer method, then teaching it to the next crew. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us (article text, 2026-10-03) / line 77 - Selected supporting excerpt: A responsible transition makes room for people to learn. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-positive-path / editorial / reviewed Label: Editorial A responsible approach starts with crews helping pick the work worth changing. People keep clear authority to accept quality and to stop unsafe work. Scope: How change should start Limitations: The paper's view. - Foundation Projects - The work above us (article text, 2026-10-03) / line 304 - Selected supporting excerpt: A responsible approach begins with crews helping identify the work worth changing and the failure cases a demonstration can miss. Decisions about quality acceptance and stopping unsafe work need clear human authority. Removing exposure to difficult tasks is valuable - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-continuity-roles / editorial / reviewed Label: Editorial Training and income should continue before paid work goes away. New roles need real demand, fair pay and a way to qualify. New people need a way to learn. Scope: What has to be built Limitations: Things to design, not guaranteed results. - Foundation Projects - The work above us (article text, 2026-10-03) / line 305 - Selected supporting excerpt: Training and income continuity belong in the deployment plan before paid work disappears. New roles in diagnosis, maintenance, quality, customer care, and machine supervision need real demand, adequate pay, and a route for people to qualify. Employers should examine how newcomers will learn when machines take on some of the repeated tasks that previously supplied practice. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-new-zero-aggressive / illustrative / reviewed Label: Scenario assumption In the Aggressive scenario the model adds no new teaching and improvement work. Support work still applies. Scope: New work in the Aggressive scenario Limitations: The model does not promise new jobs or hires. - Foundation Projects - The work above us - data README (README) / line 21 - Selected supporting excerpt: New work is zero in the Aggressive case; the model does not guarantee new occupations or hires. - Limits: The paper's notes on its data files. - Foundation Projects - The work above us - task bundles (task-bundles) / line 274 - Selected supporting excerpt: "new_human_hours_per_automated_baseline_hour": 0 - Limits: The paper's task weights and shares. The authors chose them; they are not measurements. ### cl-roles-designed / editorial / reviewed Label: Editorial New human roles must be designed and supported. They do not appear by themselves. Scope: New work for people Limitations: The paper's view. - Foundation Projects - The work above us (article text, 2026-10-03) / line 27 - Selected supporting excerpt: Those human roles must be designed and supported. Their creation is not automatic. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-gold-not-hires / illustrative / reviewed Label: Scenario assumption Gold boxes are added hours of human work, not new hires. Scope: Gold boxes in the task grids Limitations: The paper's own definition. - Foundation Projects - The work above us (article text, 2026-10-03) / line 44 - Selected supporting excerpt: Gold adds human hours around automation. It does not represent new hires. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-bls-weather / documented / reviewed Label: Official description BLS says roofers usually don't work in bad weather. Scope: Roofing seasons Limitations: Gives no count of lost days. - U.S. BLS Occupational Outlook Handbook - Roofers (Wayback capture 2026-09-20) / line 275 - Selected supporting excerpt: Roofers work outdoors in extreme temperatures, but they usually do not work during inclement weather. - Limits: Wayback copy of the BLS page; the live page could not be saved. ### cl-bls-winter / documented / reviewed Label: Official description BLS says that in northern states, roofing work may be limited in winter. Scope: Roofing seasons Limitations: Gives no count of months. - U.S. BLS Occupational Outlook Handbook - Roofers (Wayback capture 2026-09-20) / line 281 - Selected supporting excerpt: In northern states, roofing work may be limited during the winter months. - Limits: Wayback copy of the BLS page; the live page could not be saved. ### cl-machine-weather / illustrative / reviewed Label: Our assumption We assume a machine working on a roof would also stop for bad weather and winter. Scope: Roofing seasons Limitations: Our assumption. No source shows a roof machine’s weather limits. ### cl-season-example / illustrative / reviewed Label: Authored example + Derived arithmetic Example roofing season: 8 months x 20 days x 8 hours = 1,280 working hours, or 160 eight-hour days. That is about half the factory year. Scope: The example season Limitations: An example season, not data. ### cl-factory-hours / inferred / reviewed Label: Derived arithmetic Figure's BMW shift pattern (10 hours, Monday to Friday) run every weekday for a year is about 2,600 hours. Scope: The factory year Limitations: Our arithmetic: 10 hours x 5 days x 52 weeks. Figure ran for months, not a full year. Holidays are ignored. - Figure AI - Production at BMW (figure.ai, fetched 2026-10-03) / line 13 - Selected supporting excerpt: Ran 10-hour shift Monday-Friday - Limits: Figure's own page. ### cl-fixed-cost-ratio / inferred / reviewed Label: Derived arithmetic If a machine costs the same per year either way, its fixed cost per working hour grows as 2,600 divided by your working hours. Scope: Fixed cost per working hour Limitations: Our arithmetic. It assumes a yearly cost (owned or leased); a usage-based price would work differently. - Figure AI - Production at BMW (figure.ai, fetched 2026-10-03) / line 13 - Selected supporting excerpt: Ran 10-hour shift Monday-Friday - Limits: Figure's own page. - Foundation Projects - The work above us (article text, 2026-10-03) / line 357 - Selected supporting excerpt: seasonal utilization - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-cost-per-accepted / editorial / reviewed Label: Editorial The test that matters is cost per accepted task: work your customer accepts and your warranty covers. Not the purchase price or the hours a machine is switched on. Scope: The cost test Limitations: The paper's view. - Foundation Projects - The work above us (article text, 2026-10-03) / line 285 - Selected supporting excerpt: The economic threshold is cost per accepted task. Include transport, setup, supervision, downtime, maintenance, financing, rework, and warranty exposure. Compare the same roof system, access conditions, and finish quality. A low purchase price or many powered-on hours cannot answer that question. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-reading-pattern / editorial / reviewed Label: Our reading Our reading of the authors' shares: repeated work changes most. Flashing, deck repair, safety checks and handling surprises stay mostly with people. Scope: Which tasks stay with people Limitations: Our reading of shares the authors chose. - Foundation Projects - The work above us (article text, 2026-10-03) / line 273 - Selected supporting excerpt: A roofer might use a machine for a repeated installation task while continuing to diagnose damage and resolve awkward transitions. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-nailer-observed / editorial / reviewed Label: Foundation test The two clips shown here came from one AI video tool. In neither is a nail driven. In one, a screw appears where a roofing nail belongs and is left sticking up as the gun lifts away. In the other, two screws appear. Scope: The two AI clips on the roofer path Limitations: Our own look at the clips shown here. It tells you about one video tool, not about robots. ### cl-learn-from-people / self-reported / reviewed Label: Vendor-reported + Our inference Figure trains its robots on Index, a set of recordings of people at work and at home. If machines are ever trained to do roof work, we think they would learn from recordings of roofers. Scope: Recordings of work Limitations: The second sentence is our inference. Simulation and other methods are alternatives. - Figure AI - Helix 2.5 zero-shot 30-home generalization (figure.ai, fetched 2026-10-03) / line 16 - Selected supporting excerpt: we pretrained Helix 2.5 on Index, Figure's global-scale dataset of human behavior - Limits: Figure's own page. - Figure AI - Introducing Index (figure.ai, fetched 2026-10-03) / line 32 - Selected supporting excerpt: Anyone can become a Creator and start recording real tasks in their own home or workplace - Limits: Figure's own page. ### cl-index / self-reported / reviewed Label: Vendor-reported Figure says it has paid $15M to the people who record everyday tasks for it, which it calls Creators. It names tasks at home and in places like logistics centers, restaurants, factories and offices. Scope: Recordings of work Limitations: The page does not say how people are paid. - Figure AI - Introducing Index (figure.ai, fetched 2026-10-03) / line 16 - Selected supporting excerpt: We’ve paid out $15M to Creators to date that have contributed to our data - Limits: Figure's own page. - Figure AI - Introducing Index (figure.ai, fetched 2026-10-03) / line 31 - Selected supporting excerpt: To date, Creators have earned $15M. We're collecting across the full diversity of human tasks: cooking, cleaning, laundry, and other household chores at home, as well as inside businesses such as logistics centers, restaurants, factories, and offices. - Limits: Figure's own page. ### cl-index-terms-absent / editorial / reviewed Label: Our reading Figure's Index announcement (the one we checked) doesn't say how people are paid or who owns the recordings. Scope: Recordings of work Limitations: Our search of the saved page. It links to terms and conditions, which we did not review. - Figure AI - Introducing Index (figure.ai, fetched 2026-10-03) / line 16 - Selected supporting excerpt: We’ve paid out $15M to Creators to date - Limits: Figure's own page. - Figure AI - Introducing Index (figure.ai, fetched 2026-10-03) / line 32 - Selected supporting excerpt: in their own home or workplace - Limits: Figure's own page. ### cl-foundation-position / editorial / reviewed Label: Foundation position We think workers should have a meaningful part in the gains created with their experience and effort. That is a Foundation position. Whether a particular arrangement delivers it has to be tested against individual outcomes. Scope: Recordings of work and new tools Limitations: A position, not evidence. Shown once on each path. - Foundation Projects - The work above us (article text, 2026-10-03) / line 301 - Selected supporting excerpt: We think workers should have a meaningful part in the gains created with their experience and effort. That is a Foundation position - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). - Foundation Projects - The work above us (article text, 2026-10-03) / line 301 - Selected supporting excerpt: Whether a particular arrangement delivers it has to be tested against individual outcomes. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-bls-training / documented / reviewed Label: Official description BLS says trainees start by carrying equipment and material and putting up scaffolds and hoists. Within a few months they learn to measure, cut and fit. Later, they lay shingles. Scope: How roofers learn Limitations: Typical on-the-job training. - U.S. BLS Occupational Outlook Handbook - Roofers (Wayback capture 2026-09-20) / line 292 - Selected supporting excerpt: Trainees begin with tasks such as carrying equipment and material and erecting scaffolds and hoists. Within a few months, they learn to measure, cut, and fit roofing materials. Later, they lay asphalt or fiberglass shingles. - Limits: Wayback copy of the BLS page; the live page could not be saved. ### cl-entry-route / editorial / reviewed Label: Editorial If beginner tasks get easier to automate, there could be fewer ways to learn the trade. The paper calls this an idea to test, not a finding. Scope: How new roofers learn Limitations: A question to test, not a finding. - Foundation Projects - The work above us (article text, 2026-10-03) / line 280 - Selected supporting excerpt: if repeated beginner tasks become easier to automate, how will the next generation acquire expertise? A narrower entry route is a hypothesis to investigate. Experienced judgment retaining value would not, by itself, solve that problem. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-disclosure / editorial / reviewed Label: Disclosure Foundation works with roofing businesses. We have a commercial interest in this change. The paper and its model have not been peer reviewed by outside academics. Scope: Foundation Limitations: Our disclosure. - Foundation Projects - The work above us (article text, 2026-10-03) / line 432 - Selected supporting excerpt: Foundation works with roofing businesses and has a commercial interest in this transition. The article and original model have not undergone external academic peer review. - Limits: Saved article text of the paper's default view (Moderate scenario, model year 6). ### cl-bls-scale / measured / reviewed Label: Official estimate BLS says roofers held about 166,900 jobs in 2025. Median pay was $55,440 in May 2025. About 12,000 openings a year are expected over the decade, many to replace people who leave. Scope: Roofing jobs Limitations: BLS median annual wage for roofers; BLS sets who it covers. - U.S. BLS Occupational Outlook Handbook - Roofers (Wayback capture 2026-09-20) / line 268 - Selected supporting excerpt: Roofers held about 166,900 jobs in 2025. - Limits: Wayback copy of the BLS page; the live page could not be saved. - U.S. BLS Occupational Outlook Handbook - Roofers (Wayback capture 2026-09-20) / line 228 - Selected supporting excerpt: The median annual wage for roofers was $55,440 in May 2025. - Limits: Wayback copy of the BLS page; the live page could not be saved. - U.S. BLS Occupational Outlook Handbook - Roofers (Wayback capture 2026-09-20) / line 231 - Selected supporting excerpt: About 12,000 openings for roofers are projected each year, on average, over the decade. Many of those openings are expected to result from the need to replace workers who transfer to different occupations or exit the labor force, such as to retire. - Limits: Wayback copy of the BLS page; the live page could not be saved. ### cl-generated-not-evidence / editorial / reviewed Label: Project decision Every generated clip and image is labelled. Robots appear only on a practice roof. The nailer clips are shown only as failures. Nothing generated supports a fact. Scope: Every generated clip and image on this page Limitations: Our rule for this page. ## Continue Map My Exit: https://foundationproject.net/schedule ## Build identity Project: roof-robots Run: 20261007T151842621Z-4812bf67 Content fingerprint: 39ef8142ed7c48b2882ed397fe47da7bceaac945a48eb7b6da8fb4a03f2f1283 Mode: public Fonts: Newsreader (SIL Open Font License 1.1; Copyright 2020 The Newsreader Project Authors (http://github.com/productiontype/Newsreader)); Inter (SIL Open Font License 1.1; Copyright 2016 The Inter Project Authors (https://github.com/rsms/inter)) Music: generated with Higgsfield Sonilo. Music is optional: it stays off until the visitor turns sound on, and the page says everything without it. Publication: local artifact; this file does not establish a live deployment. A download records an action, not later use, learning or business value.