Anthropic published survey results on 16 October 2026 comparing its own staff with US adults, and found the two groups largely agree on what AI can do today but split on how fast it will improve: “91% of staff expect AI to be able to make a Nobel-level discovery by 2030, compared to 39% of US adults.”
The gap carries into work and the economy. Anthropic says staff expect AI to be able to do 75% of their current work by 2030, against 28% for US workers, and that the median staff member’s answers imply GDP growth of 17.1% in 2030, against 5.4% for the median US adult.
Business Engineer Video · AI ADOPTION
A short preview of a Business Engineer video on the state of enterprise AI adoption, made with Ramp Economics Lab. It covers adoption data in general and says nothing about Anthropic’s survey.
The key insight: As we read it, the survey separates two questions that usually get blurred: Anthropic says US adults and its staff largely agree on what AI can do today, and the gap opens on how fast it improves and spreads, with 91% of staff against 39% of US adults expecting a Nobel-level AI discovery by 2030.
The note, “The AI expectations gap”, is by Szymon Sacher, Tess Cotter, Charles I. Jones, Anton Korinek and Peter McCrory of the Anthropic Institute. It compares two surveys fielded in August 2026: a nationally representative sample of US adults and a poll of Anthropic staff.
According to the appendix, Morning Consult fielded the general-public survey between August 11 and 23, 2026, and after removing repeat interviews it kept 10,980 respondents, weighted to the US adult population. The staff poll ran from August 9 to 12, 2026, and drew 1,334 respondents; Anthropic says staff responses are unweighted.
Anthropic is explicit that the staff numbers are personal views. On its hiring question it says the answers “are the personal expectations of staff members in an internal poll and do not represent Anthropic’s plans”.
Per Anthropic’s research note “The AI expectations gap” (16 October 2026): 91% of Anthropic staff vs 39% of US adults expect AI to be able to make a Nobel-level discovery by 2030; 59% of staff vs 26% of US adults who use AI at work say a task takes a quarter of the time or less; staff expect AI to be able to do 75% of their own work by 2030 on average, US workers 28%.
What Was Posted on X
Surveyed Anthropic staff’s answers imply 17.1% GDP growth in 2030
Lisan al Gaib (@scaling01) posted on 16 October that the surveyed staff’s answers imply 17.1% GDP growth in 2030, the median figure Anthropic’s note reports from its economic model.
Over half of surveyed Anthropic staff believe AI will be able to make Nobel Prize-worthy discoveries by 2027
In a follow-up post he wrote that over half of surveyed Anthropic staff believe AI will be able to make Nobel Prize-worthy discoveries by 2027. The note’s text gives the Nobel figure for 2030 (91% of staff); the 2027 share is his reading of the note’s Figure 1 chart.
Where the Two Groups Split
Anthropic says the groups “largely agree about what AI can do today” and rank eight tasks, from routine emails to a Nobel-level discovery, in nearly the same order of difficulty. They split on pace.
About a quarter of both groups think AI can already design, conduct and publish scientific studies. “But by next year, 9 in 10 staff expect this to be feasible for AI as compared to just 38% for the general public.”
Between 19% and 40% of US adults, depending on the task, say AI will never be able to do it, against at most 2% of staff. Use matters too: among US adults who use AI at work regularly, 56% expect AI to manage at least seven of the eight tasks by 2030, against 41% of all US adults and 96% of staff.
Time Saved and Own Jobs
On productivity, 85% of staff say a task suited to AI takes half the time or less with it. Among US adults who use AI at work, 26% say such a task takes a quarter of the time or less, against 59% of staff.
On their own jobs, staff put the share of their current work AI could do by 2030 at 75% on average, against 28% for US workers. Only 1 in 20 US workers say AI could do 80% or more of their work hours by then; Anthropic says over half of staff expect at least 80%.
Both groups link own-job exposure to unemployment the same way, the note says. US workers who expect AI to do 80% or more of their work forecast an average 2030 unemployment rate of 12.2%, and across US workers each ten more points of affected work goes with 0.4 points more forecast unemployment. Anthropic finds “the same gradient” among staff.
From Answers to a GDP Number
Anthropic ran each respondent’s answers on capabilities, diffusion, automation, productivity and re-employment through the model in its paper Economic Scenarios for Transformative AI, using 3,259 US adults and 1,093 staff who answered every required question.
“The median US adult’s answers imply GDP growth of 5.4% in 2030, according to the model. The median staff member’s answers imply GDP growth accelerating to 17.1% by 2030.” Anthropic adds that the distributions overlap widely.
A third group, 23,075 visitors who took Anthropic’s online scenario quiz in the 48 hours after it launched on September 9, 2026, landed in between: their median answers imply 12.5% growth, the note says. In its conclusion, Anthropic writes that the staff expectations imply that “by 2030, GDP is doubling in size roughly every four and a half years”.
Inside the Lab
Anthropic also asked staff what share of the work of building an early-2026 frontier model AI does. “Among surveyed Anthropic staff, the median answer today is 51%, 80% by the end of 2027 and 95% by the end of 2030.” Staff on research teams put today’s share at 60% on average, against 53% for other respondents.
Lisan al Gaib (@scaling01) posted on 16 October that “Over half of surveyed Anthropic staff believe AI will be able to make Nobel Prize-worthy discoveries by 2027”. The 2030 figure is in Anthropic’s text; the 2027 share appears only in the note’s Figure 1 chart, so that reading is his.
The Structural Read
As we read it, exposure moves expectations: Anthropic says US adults who use AI at work regularly already sit closer to staff, with 56% expecting AI to manage at least seven of eight tasks by 2030, against 41% of all US adults and 96% of staff.
As we read it, the disagreement is about inputs, not reasoning: Anthropic says both groups link the share of their own work AI can do to higher forecast unemployment along the same gradient.
As we read it, the GDP gap is the survey gap passed through a model: Anthropic’s economic-scenarios model turns the median staff answers into 17.1% growth in 2030 and the median US adult’s into 5.4%.
Anthropic Institute, The AI expectations gap (16 October 2026)
“Most Anthropic staff already work alongside AI that does much of their job, whereas few US workers do.”
Three Implications
COMPANIES PLANNING AI ADOPTION Anthropic says US adults name customers preferring a person, data privacy and security, reliability, and laws or liability as the biggest obstacles to AI at work in 2030.
WORKERS AND LABOR-MARKET WATCHERS Anthropic says US workers who expect AI to do 80% or more of their work forecast 12.2% unemployment in 2030, and staff show the same gradient at higher exposure.
ANYONE READING THE STAFF NUMBERS Anthropic says the staff answers come from an internal, unweighted poll of its own employees and are personal views, not company forecasts or plans.
The framework describes two curves, what a technology can do and what users can actually do with it, and says: “They are not the same curve”.
As we read it, Anthropic’s survey measures a version of that gap from the user side: staff who work with frontier models every day expect far faster progress than US adults who mostly do not. The framework predates the survey and does not discuss it.
Per Anthropic’s research note (16 October 2026, Figure 7): each respondent’s answers run through the model in Economic Scenarios for Transformative AI imply 2030 GDP growth of 5.4% for the median US adult, 12.5% for the median website-quiz visitor and 17.1% for the median Anthropic staff member.
What Is Not Established
The staff sample is Anthropic’s own employees, polled internally and unweighted, and Anthropic says those answers are personal views, not company forecasts. The GDP figures are model outputs from survey answers, not forecasts by Anthropic.
We read the research note and its appendix on 16 October 2026. We did not contact Anthropic or Morning Consult and did not see the full questionnaire.
Business Engineer Framework
The Product Overhang Doctrine
A Business Engineer framework on the gap between what a technology can do and what its users can do with it.
Anthropic said on 16 October 2026 that 91% of its staff, against 39% of US adults, expect AI to be able to make a Nobel-level discovery by 2030. It said staff expect AI to be able to do 75% of their own work by then, against 28% for US workers.
Run through Anthropic’s economic model, the median staff member’s answers imply 17.1% GDP growth in 2030 and the median US adult’s 5.4%. Anthropic notes that US adults who use AI at work already hold views closer to the staff’s.
A note on sourcing. We read in full the Anthropic Institute research note “The AI expectations gap” (16 October 2026) and its appendix on 16 October 2026; every quotation and figure about the survey comes from those texts, and the quoted X post is by Lisan al Gaib. The staff figures come from an internal poll of Anthropic employees. We did not contact Anthropic or Morning Consult. The references to our earlier reporting link to our own pieces.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur, HackerNoon and Search Engine People. Find him on LinkedIn and Substack. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
