Here is a 3-paragraph summary of the blog post for mathopen.com:
Scott Aaronson, a researcher at UT Austin, sat down for a wide-ranging 90-minute in-person conversation on the "Knowmads" podcast, recorded right in his office. The relaxed setting gave the discussion an informal and candid feel, with Aaronson's blackboard visible in the background, featuring a mysterious bulleted list including terms like "ARC," "Scalable Oversight," and "Models" that even he could not explain after the fact.
The podcast covers topics at the intersection of science and artificial intelligence, areas that Aaronson has thought deeply about through his work in theoretical computer science and quantum computing. The long-form format allowed for a thorough and thoughtful exploration of ideas, the kind of deep dive that shorter interviews rarely permit.
For anyone interested in hearing a leading researcher think out loud about AI, its challenges, and its connections to fundamental science, this episode is well worth the listen. You can stream it directly through the link provided on the blog, and if that does not work, an alternative link is also available to make sure you do not miss out.