Dwarkesh, I think you’re about to pay for the wrong unit of work.
Scouting is a conversion problem: turning an infinite ocean of expertise into a small stream of guests who are obviously banger in conversation. Most people can talk about their paper. Very few can do the Reich/Kimmel thing: field map + hot takes + real synthesis, without turning into TED fog or PR.
So: don’t pay for “names.” Pay for decision-ready packets.
If I were you I’d standardize the output to a one-page dossier:
• 1–2 clips with timestamps (must include Q&A)
• three claims they can defend conversationally
• 8–10 question stems (including “what would change your mind?”)
If a scout can’t produce that, they don’t actually have a lead, they have a vibe.
Also: you can lean on AI way more here. AI can do the unsexy part cheaply (search seminar archives, pull transcripts, find the best moments, draft the dossier). Humans should spend their time on taste: “is this person alive,” “do they range,” “are incentives clean.”
One more compounding trick: every great guest should yield five more. Ask them who gave the best talk they’ve seen in two years, who they disagree with, who’s underrated. That flywheel beats cold scouting.
Net: you probably don’t need to spend $2K/week to get better guests. You need a tighter unit of output + an AI-assisted pipeline. Happy to share an example dossier if useful.
You should interview Robert Sapolsky - Stanford professor/author known for explaining how stress, biology and evolution shape human behavior with a strong emphasis on questioning the concept of free will
I think Frank Dikötter would be a great interview partner for you - it would also perfectly fit the episodes you had with Sarah Paine. He has a new book coming out next week on the rise of the Communist Party in China.
Dwarkesh already invited Razib Khan, Charles Muray, Garret Jones, Joseph Henrich and Gregory Clark years ago (same genre). It might be some the first episodes in his youtube history.
(The first ones were the GMU econ crew with Brian Caplan, Tyler Cowen Robin Hanson etc...)
Ruxandra Teslo could work well for biotech - has pretty wide ranging takes including the questions you mentioned, and AFAIK not affiliated with any company. https://substack.com/@ruxandrabio
How will sort between all the recuring demands for Joscha Bach, Karl Friston and Mike Israetel? ^^
Btw the go to guy for chinese economic history, the great divergence, innovation and the current Thucidide trap is Yasheng Huang from MIT. His work on historical chinese innovation is a tour de force.
Sounds like Sarah Paine might enjoy a conversation with Yasheng Huang. Asian history is one of her specialties. & economics is certainly related to grand strategy.
You should have more repeat guests. Bring em back a year later, you'll have plenty to talk about.
Dwarkesh, I think you’re about to pay for the wrong unit of work.
Scouting is a conversion problem: turning an infinite ocean of expertise into a small stream of guests who are obviously banger in conversation. Most people can talk about their paper. Very few can do the Reich/Kimmel thing: field map + hot takes + real synthesis, without turning into TED fog or PR.
So: don’t pay for “names.” Pay for decision-ready packets.
If I were you I’d standardize the output to a one-page dossier:
• 1–2 clips with timestamps (must include Q&A)
• three claims they can defend conversationally
• 8–10 question stems (including “what would change your mind?”)
• a tiny prep curriculum (3 reads, 2 talks, 5 terms)
If a scout can’t produce that, they don’t actually have a lead, they have a vibe.
Also: you can lean on AI way more here. AI can do the unsexy part cheaply (search seminar archives, pull transcripts, find the best moments, draft the dossier). Humans should spend their time on taste: “is this person alive,” “do they range,” “are incentives clean.”
One more compounding trick: every great guest should yield five more. Ask them who gave the best talk they’ve seen in two years, who they disagree with, who’s underrated. That flywheel beats cold scouting.
Net: you probably don’t need to spend $2K/week to get better guests. You need a tighter unit of output + an AI-assisted pipeline. Happy to share an example dossier if useful.
You should interview Robert Sapolsky - Stanford professor/author known for explaining how stress, biology and evolution shape human behavior with a strong emphasis on questioning the concept of free will
No. Sapolsky peddles unreplicable findings
I'd second this -- and that his free will stuff hasn't really added anything beyond what Sam Harris has already said
How about Alex Wellerstein?
Short bio: Historian of science and nuclear weapons and a professor at the Stevens Institute of Technology
Creator of NUKEMAP: https://nuclearsecrecy.com/nukemap/
WIRED video - "Nuclear Historian Answers Nuclear War Questions": https://www.youtube.com/watch?v=CJTA2OinEHw
Full disclosure... I went to highschool with him.
I think Frank Dikötter would be a great interview partner for you - it would also perfectly fit the episodes you had with Sarah Paine. He has a new book coming out next week on the rise of the Communist Party in China.
Frank Dikötter, author of Mao‘s Great Famine and many other excellent books on China.
Highly polymathic and somewhat in the David Reich 'lane': Razib Khan!
https://www.razibkhan.com/archive?sort=top
I'm not sure how many scientists know as much history as Razib does, nor how many historians know as much science.
Dwarkesh already invited Razib Khan, Charles Muray, Garret Jones, Joseph Henrich and Gregory Clark years ago (same genre). It might be some the first episodes in his youtube history.
(The first ones were the GMU econ crew with Brian Caplan, Tyler Cowen Robin Hanson etc...)
Ruxandra Teslo could work well for biotech - has pretty wide ranging takes including the questions you mentioned, and AFAIK not affiliated with any company. https://substack.com/@ruxandrabio
You must get Prof. Steve Keen for economics!
Looks like this guy is imitating you on Substack.
https://substack.com/@dwarkeshgsjsbsjsvsjsnsjz?r=rkcg&utm_medium=ios&utm_source=profile&shareImageVariant=image
Glen Weyl is quite creative and polymathic, and has a unique perspective on how AI should be developed to spread the benefit to society at large.
Good to know you can't just create AI Agent Scouts
This job posting might actually be: paying humans to curate the best possible AI Agent Scouts
How will sort between all the recuring demands for Joscha Bach, Karl Friston and Mike Israetel? ^^
Btw the go to guy for chinese economic history, the great divergence, innovation and the current Thucidide trap is Yasheng Huang from MIT. His work on historical chinese innovation is a tour de force.
Sounds like Sarah Paine might enjoy a conversation with Yasheng Huang. Asian history is one of her specialties. & economics is certainly related to grand strategy.
David Eagleman - AI x Neuroscience. Stanford professor and author.
Alice Evans. 10000 years of patriarchy.
Would love to see you interview Yuval Noah Harari