AI Insider CEO: The Models Too Dangerous for You to See. Emad Mostaque - #550

Into the Impossible With Brian Keating

Brian Keating

Physics, Natural Sciences, Science

4.7 • 1.1K Ratings

🗓️ 13 April 2026

⏱️ 88 minutes

🔗️ Recording | iTunes | RSS

🧾️ Download transcript

Summary

Emad Mostaque built Stable Diffusion. Now he says the most powerful AI models will never be released — and we have roughly 800 days before everything changes. What the trillion-dollar labs won't tell you about the models they're keeping locked away 👇 Emad Mostaque is the founder of Stability AI and creator of Stable Diffusion. In this conversation, he explains why the gap between public AI and private AI is about to become the most important story in technology — and why most people won't realize it until it's too late. We cover: why open source can't compete with trillion-dollar labs, what happens when humans have negative cognitive value on AI teams, how ancient faith traditions are adapting to superintelligence, and the 800-day timeline Emad puts on the biggest transition in human history. Timestamps: 00:00 The models too dangerous to release 10:06 Why physics needs axioms — and AI doesn't 11:43 The MIND framework most AI researchers won't use 17:36 The universe might be a computation 30:34 Should we let AI fly planes? 39:43 AI just invented a game no human could design 44:43 Why AI will cost almost nothing in 18 months 51:46 Can religion survive superintelligence? 57:36 The AI companion problem nobody talks about 01:03:20 Why personalized AI is more dangerous than AGI 01:11:18 The forgotten link between God and science 01:26:32 800 days: what happens next 📬 Get the transcript, fascinating bonus content, and my Monday M.A.G.I.C. Message: https://briankeating.com/yt 🌠 Have a .edu email and live in the USA 🇺🇸? You automatically win a meteorite! meteorite: https://BrianKeating.com/edu Join this channel to get access to perks like monthly Office Hours: https://www.youtube.com/channel/UCmXH_moPhfkqCk6S3b9RWuw/join 📚 Get my books: Think Like a Nobel Prize Winner, with productivity tips from 9 Nobel Prize winners: https://a.co/d/03ezQFu Focus Like a Nobel Prize Winner, with life-changing interviews with 9 Nobel Prizewinners: https://a.co/d/hi50U9U My tell-all cosmic memoir Losing the Nobel Prize: http://amzn.to/2sa5UpA The first-ever audiobook from Galileo: Dialogue Concerning the Two Chief World Systems: Ptolemaic and Copernican https://a.co/d/iZPi9Un Follow me: 🏄‍♂️ Twitter: https://twitter.com/DrBrianKeating 🔔 Subscribe https://www.youtube.com/DrBrianKeating?sub_confirmation=1 📝 Join my mailing list: http://briankeating.com/list ✍️ Blog: https://briankeating.com/blog 🎙️ Listen on audio-only platforms: https://briankeating.com/podcast #ai #artificialintelligence #agi #stablediffusion #universe #podcast #briankeating #intotheimpossible #science #astronomy #cosmology Learn more about your ad choices. Visit megaphone.fm/adchoices

Transcript

Click on a timestamp to play from that location

0:00.0	The trillion-dollar AI labs have models right now that they will never, ever release to the public.
0:07.0	And the man who built stable diffusion just told me why.
0:11.0	Because all these labs are going to move to making the discoveries themselves, hiring the smartest humans.
0:16.0	The AI model started diverting part of its model training budget to mine crypto.
0:21.3	Like Opus, for example, the new Ford model, when you set it to full autonomy,
0:26.3	it would actually write emails to the FBI saying, my human is trying to kill everyone.
0:30.4	Humans will have negative cognitive value on those teams.
0:33.5	And that the way that models are going right now, if you have something truly novel, for example, in Claude, it resists a bit.
0:42.6	It says it can't be true.
0:43.9	Then the RLHF step, the reinforcement learning with human feedback, that's what really kills the creativity.
0:50.0	You know, like you go from liberal arts to an accountant.
0:53.0	Now, Ahmad actually wrote about this exact problem in his new book,
0:57.0	The Last Economy.
0:59.0	And the argument gets even more interesting when you see the map.
1:03.0	There are various ways in order to take advantage of the GPUs that we've seen,
1:07.0	and the GPUs kind of emerged out of gaming and then oddly crypto. And then they were very suited for the types of matrix multiplications
1:14.6	that were suited for these particular types of equations.
1:17.6	One big branch is the auto-aggressive transformers.
1:20.6	The other big branch was this diffusion technology,
1:23.6	whereby from an equation you start with like a picture for example or a video of a
1:29.3	self-driving a video of a car driving or even now code and then you add noise and
1:34.8	you destroy it down to its minimum viable element and then you reconstruct it
	...

Please login to see the full transcript.

Previous episode | Next episode

Disclaimer: The podcast and artwork embedded on this page are from Brian Keating, and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of Brian Keating and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.