Episode 126: Using AI to Improve Science (with Paul Litvak)

Paul Litvak joins the show to talk about how AI tools can help us measure research quality and assess evidence in the scientific literature. His first project is a way to extract test statistics and p-values from papers automatically, with no manual coding needed. We also talk about Paul's non-profit dedicated to improving the reliability of scientific research, the legendary judgment and decision making scholar Robin Dawes (whose entirely algorithmic approach to graduate student selection once went terribly awry), and Paul's exit from academia. Plus, Yoel reveals a shameful secret about his use of AI.

Special Guest: Paul Litvak.

Links:

What If Everyone Knew Which Science to Trust?

evidence.guide

The Robyn Dawes Institute for the Improvement of Science

Why are so many professors conservative? - by Paul Bloom

Science is a strong-link problem - by Adam Mastroianni