Will AI make a Nobel-level discovery by 2030? Most Americans say no. AI builders say yes.
Anthropic asked US adults and its own staff the same questions. 39% of US adults expect a Nobel-level AI discovery by 2030. Among Anthropic staff, it was 91%.
How fast will AI get better? It depends on who you ask.
Anthropic, the company that makes the Claude chatbot, asked two groups the same questions. One group was US adults. The other group was Anthropic's own staff, the people who build AI every day. The survey ran in August 2026. The results came out on October 16.
The biggest gap is about science. By 2030, 39% of US adults expect AI to make a Nobel-level discovery. Among Anthropic staff, 91% expect it.
Many people think AI will never do some jobs. Between 19 and 40% of US adults said that, depending on the task. Among staff, it was at most 2%.
The two groups also see their own jobs very differently. On average, staff said AI will be able to do 75% of their work hours by 2030. US workers said 28%. Only 1 in 20 US workers said AI could do 80% or more of their work.
What could slow AI down at work? US adults picked people more than machines. The top answer was "Customers preferring to deal with a person". Next came "data privacy and security concerns".
The two groups do agree on one thing. In both groups, people who expect AI to do more of their own work also expect more people to lose jobs.
So who is right? Nobody knows yet. But the gap tells us something. The people closest to AI expect much bigger changes, and much sooner, than everyone else.
Open it and check the claim yourself. That is the point.