Navigate Select ESC Close

Roman Yampolskiy: AI Can't Be Controlled. We Need to Pause Now!

2026-06-21 Science & Technology
184
15
2
Dr Brian Keating
Dr Brian Keating
417.0k subscribers

Unlock all features

FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.

Description

Roman Yampolskiy spent two decades trying to prove superintelligence can be controlled. He couldn't — and now says the proof runs the other way: not difficult, mathematically impossible. If he's right, we're building the one machine we can never switch off, and the people building it are racing to do it faster. Roman is a professor of computer science at the University of Louisville and one of the earliest researchers in AI safety. We cover: why you're the squirrel in the human-AI intelligence gap, what the halting problem really says about predicting a smarter mind, why he calls all current AI safety "security theater," and the one button he'd press if he could. In theory communism works too. In practice, no human verifies a billion-page proof. CHAPTERS 00:00 The mind we can't switch off 01:34 You're the squirrel in this story 04:08 The thought AI may never have 08:18 We ran the test. It failed. 12:22 The trap we already locked ourselves in 16:10 Why AI still can't crack physics 21:54 Why toddlers are safe and teenagers aren't 23:06 The proof you can't predict a smarter mind 25:58 No system can vouch for itself 28:18 Does regulation just protect the oligarchs? 33:28 Is your mind more than a machine? 39:00 Do we owe AI anything? 43:00 The monster behind the smiley face 46:22 A perfect universe built just for you 49:50 What if we're the alien probes? 55:02 The one kind of AI that should scare you 58:22 The movement begging labs to stop 59:58 What do you tell your kids now? 01:07:58 The question he'd put to Sam Altman 01:15:48 The debate he's dying to have 01:19:50 What he'd tell his 20-year-old self 📬 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: https://BrianKeating.com/edu 🔔 Subscribe: https://www.youtube.com/DrBrianKeating?sub_confirmation=1 🎯 Support Into the Impossible on Patreon — get my weekly M.A.G.I.C. Message, unfiltered bonus content, and live monthly Office Hours with me: https://www.patreon.com/drbriankeating ⭐ Join this channel for perks, monthly Office Hours, and your name in the Member Roster at the end of every episode: https://www.youtube.com/channel/UCmXH_moPhfkqCk6S3b9RWuw/join Featured Guest — Roman Yampolskiy The Roman Forum (YouTube): https://www.youtube.com/c/RomanYampolskiy?sub_confirmation=1 Twitter/X: https://x.com/romanyam AI: Unexplainable, Unpredictable, Uncontrollable: https://www.romanyampolskiy.com/books/ 📚 My books: Losing the Nobel Prize (memoir): http://amzn.to/2sa5UpA Think Like a Nobel Prize Winner: https://a.co/d/03ezQFu Focus Like a Nobel Prize Winner: https://a.co/d/hi50U9U Galileo's Dialogue (first-ever audiobook): https://a.co/d/iZPi9Un 🌐 More: 🏄‍♂️ Twitter: https://twitter.com/BrianKeating 📚 Substack: https://briankeating.substack.com/ss ✍️ Blog: https://briankeating.com/blog 🎙️ Audio-only: https://briankeating.com/podcast #intotheimpossible #briankeating #science #physics #astronomy #cosmology #podcast #universe #AIsafety

Top Comments (10)

@DrBrianKeating 2026-06-21

Roman Yampolskiy started his career trying to prove AI could be made safe — and ended up proving the opposite. Not "hard." He says mathematically impossible. 📬 Transcript + my Monday M.A.G.I.C. Message → briankeating.com/yt Steelman for the other side: David Deutsch says humans are universal explainers, so no problem — including control — is forever closed to us. Roman's reply is that you can't predict, verify, or test a mind running non-deterministic states you can't even survey. The 3 moments I'd re-watch: ▸ 1:34 — Why you're the squirrel in this story ▸ 23:06 — The halting problem comes for superintelligence ▸ 1:07:58 — The button he'd actually press If control really is unsolvable, is a permanent ban rational — or do we build it anyway? Pick one and defend it. 👇

8 1 replies
@Johnnynyny 2026-06-21

Make it Stop! If Humans don’t let autocomplete word-guessing software control: weapons, vehicles, economies, facts or other important stuff, maybe we won’t die young- Super Intelligence 😂😂😂😂

3
@sarcasmunlimited1570 2026-06-21

Yampolskiy, Yudkowsky, and Hinton know they aren't as smart as today's frontier AIs. That's why they think ASI is close. Meanwhile, LLMs can never achieve AGI, and other methods are still in their infancy.

2 4 replies
@Thomas-gk42 2026-06-21

He has a point with his animal analogy.

2
@merlepatterson 2026-06-22

As it's said, we might be closer to zero than we are to infinity, but if there is no such thing as zero nor infinity, then what are we really closer to?

1 3 replies
@Ajax-0137 2026-06-22

Fear mongering is out of control

1
@chainboxchainbox6465 2026-06-22

No they take patterns that someone created my friend.😊

1
@SERGEX42069 2026-06-24

Why pause? Full acceleration. Get rid of all of us asap. Thats the goal. Why stand in the way?

0
@chainboxchainbox6465 2026-06-22

Ahhhh I'm laughing and the answers for any kind of proof no matter the circumstances would not need more than a chapter......😊

0
@Synapscience 2026-06-24

Stanley Kubrick started thinking about AI safety in 1968 and gave us the Tragedy of HAL as a warning. I started thinking about AI Safety in 1996 when I created the world's first AUTONOMOUS AGENTS FOR THE INTERNET and called my AI software AUTONOMY. Roman Yampolskiy has no idea what intelligence is or what IQ is, he thinks that he thinks with his brain and he thinks AI will replace 99% (yes 99%) of jobs. Try telling this to STRIPPERS, CARDINALS, BANK ROBBERS, ARTISTS, SUMO WRESTLERS, COMEDIANS, ELVIS IMPERSONATORS, MONKS, SOMMELIERS, ASSASSINS....I could go on. Dr Keating...Noblesse Oblige....be careful who or what you endorse. Dr Yampolskiy, as Mark Twain would say, it is better to remain silent and be suspected an ignorant fool than to open one's mouth and remove all doubt.

0

Unlock the Data Inside
Turn Videos into Knowledge

  • Get FREE 10/day: transcripts, summaries, chats
  • Chat with videos, export text & PDF
  • $1 free API credit for RAG, chatbots & research

Free forever plan • All features unlocked

App screenshot