We ran a randomized controlled trial to see how much AI coding tools speed up experienced open-source developers. The results surprised us: Developers thought they were 20% faster with AI tools, but they were actually 19% slower when they had access to AI than when they didn't.

Are you an early-career cognitive neuroscientist, looking to contribute to a new line of research? The Frontlab at Paris Brain Institute is seeking a postdoctoral researcher to work on a state-of-the-art creativity neuroscience project - offres.institutducerveau-icm.org/fr/jobs/1956-75

Really interesting new 𝔊𝔴𝔢𝔯𝔫 essay: LLM Daydreaming - Proposal of how default mode networks for LLMs are an example of missing capabilities for search and novelty Btw, I know it's a bit cringe to delight in, but if you had told 19 year old me that a Gwern essay would open

Creativity has been linked to better learning—but why? We find that both rely on associative thinking: the ability to form new connections. nature.com/articles/s4153…

A comprehensive and accessible overview of LLMs for psychological assessment journals.sagepub.com/doi/full/10.11…

Surprised not to see much psychological research on how AI impacts creativity in kids—does it help or hurt their imagination? Does it make them more or less creative than adults? Maybe I missed it, or it’s still in the works. I know research with kids takes time 😅

I talked to NPR about AI’s impact on creativity, alongside two art professors. We covered a lot of ground—maybe even found some middle ground—on the benefits and costs of AI in creative work. ctpublic.org/show/where-we-…

More evidence of LLM homogenization: college admissions essays submitted to a highly selective university were more similar in their ideas and themes after ChatGPT came out, especially for applicants from backgrounds outside the cultural mainstream. osf.io/preprints/psya…

Researchers filmed 6 mathematicians at the blackboard. Just before their “aha!” moments, their writing and gestures grew more erratic—suggesting that insights may have detectable "early warning signs" in behavior. pnas.org/doi/full/10.10…