Read the full transcript here.

Is AI going to ruin everything? What kind of AI-related dangers should we be most worried about? What do good institutions look like? Should designing better institutions be a major priority for modern civilizations? What are the various ways institutions decay? How much should we blame social media for the current state of our institutions? Under what conditions, if any, should the flow of information be regulated? What are some of the lesser-known kinds of AI disalignment? What actions should we take in light of the lack of consensus about AI?

Gabe Alfour has a background in theoretical computer science and has long been interested in understanding and tackling fundamental challenges of advancing and shaping technological progress. Fresh out of university, he developed a new programming language and founded a successful French crypto consultancy. Gabe has long had an interest in artificial intelligence, which he expected to be a major accelerator of technological progress. But after interacting with GPT-3, he became increasingly concerned with the catastrophic risks frontier AI systems pose, and decided to work on mitigating them. He studied up on AI and joined online open-source AI community EleutherAI, where he met Connor Leahy. In 2022, they co-founded Conjecture, an AI safety start-up. Gabe is also an advisor with ControlAI, an AI policy nonprofit. Email Gabe at ga@conjecture.dev, follow him on Twitter / X at @Gabe_cc, or read his writings on his blog at site.cognition.cafe.

Staff

• Spencer Greenberg — Host / Director
• Josh Castle — Producer
• Ryan Kessler — Audio Engineer
• Uri Bram — Factotum
• WeAmplify — Transcriptionists
• Igor Scaldini — Marketing Consultant

Music

• Broke for Free
• Josh Woodward
• Lee Rosevere
• Quiet Music for Tiny Robots
• wowamusic
• zapsplat.com

Affiliates

• Clearer Thinking
• GuidedTrack
• Mind Ease
• Positly
• UpLift

Read the full transcript here.

What defines a cult? Is there such a thing as a good cult? Do Clearer Thinking's tools actually help people? Why does Clearer Thinking share its tools for free for everyone to use? How legitimate is the research Clearer Thinking does to create its tools? Is that research too reliant on self-report? Do Clearer Thinking's tools focus too much on the average person and fail to account for significant variance among people? Should AI companies be required to create and release text watermarking tools? Should smart, knowledgeable people speak out more? Would the average person think (without priming or knowledge of the discourse around it) that Elon Musk's gesture at the inauguration was a Nazi salute? Does Spencer sometimes coin new terms where useful terms already exist? Does Spencer think that everyone should adopt valuism, his life philosophy? Is magic real? What critiques have stuck with Spencer over the years and shaped his work?

Staff

• Spencer Greenberg — Host / Director
• Josh Castle — Producer
• Ryan Kessler — Audio Engineer
• Uri Bram — Factotum
• WeAmplify — Transcriptionists
• Igor Scaldini — Marketing Consultant

Music

• Broke for Free
• Josh Woodward
• Lee Rosevere
• Quiet Music for Tiny Robots
• wowamusic
• zapsplat.com

Affiliates

• Clearer Thinking
• GuidedTrack
• Mind Ease
• Positly
• UpLift

Read the full transcript here.

NOTE: The video version of this conversation is available on YouTube: https://youtu.be/hWNknrc23Fo

In light of the replication crisis, should social scientists try to replicate every major finding in the field's history? Why is human memory so faulty? And since human memory is so faulty, why do we take eyewitness testimony in legal contexts so seriously? How different are people's experiences of the world? What are the various failure modes in social science research? How much progress have the social sciences made implementing reforms and applying more rigorous standards? Why does peer review seem so susceptible to importance hacking? When is observation more important than interpretation, and vice versa? Do the top journals contain the least replicable papers? What value do Freud's ideas still provide today? How useful are neo-Freudian therapeutic methods? Should social scientists run studies on LLMs? Which of Paul's books does ChatGPT like the least?

Paul Bloom is Professor of Psychology at the University of Toronto, and Brooks and Suzanne Ragen Professor Emeritus of Psychology at Yale University. Paul Bloom studies how children and adults make sense of the world, with special focus on pleasure, morality, religion, fiction, and art. He has won numerous awards for his research and teaching. He is past-president of the Society for Philosophy and Psychology, and co-editor of Behavioral and Brain Sciences. He has written for scientific journals such as Nature and Science, and for popular outlets such as The New York Times, The Guardian, The New Yorker, and The Atlantic Monthly. He is the author of seven books, including his most recent, Psych: The Story of the Human Mind. Find more about him at paulbloom.net, or follow his Substack.

Staff

• Spencer Greenberg — Host / Director
• Josh Castle — Producer
• Ryan Kessler — Audio Engineer
• Uri Bram — Factotum
• WeAmplify — Transcriptionists
• Igor Scaldini — Marketing Consultant

Music

• Broke for Free
• Josh Woodward
• Lee Rosevere
• Quiet Music for Tiny Robots
• wowamusic
• zapsplat.com

Affiliates

• Clearer Thinking
• GuidedTrack
• Mind Ease
• Positly
• UpLift