These impacts were measured before practical agents (like Claude Code) and companies are still early in figuring out how to incorporate AI into their workflows.
Ethan Mollick
None of the things people hate about AI are here: it is not trained on copyright information, it is not controlled by companies, it is not using a lot of resources, it is small and fun. It is also a genuinely interesting potential direction for research. And yet I have to keep blocking mad people.
The average American worker using AI reports time savings of 6%, or 2.5 hours in a work week. Those are similar to the UK & Netherlands, slightly more than other EU countries.
There some early, non-causal, signs that this is translating into gains in productivity growth www.nber.org/papers/w34995
Ethan Mollick
Ethan Mollick
My prediction for the latest trend in academic buildings: Faraday cage testing halls (including bathrooms) with no signal for assessment.
(I am doing the exact same thing, of course, because it is humans who judge my work and who I want to influence with my writing, but I want to start putting out .md versions as well))
The fact that every scientific paper in 2026 is still uploaded only as fully formatted PDFs to academic archive sites that often limit downloads tells you everything you need to know about how quickly the scientific system is adjusting to the potential of AI to accelerate science & help discovery.
