Roman Yampolskiy

Date: 2025-07-03 | Duration: 2h 18m

Episode Overview

AI safety researcher Roman Yampolskiy sits across from Joe to sound the alarm on artificial general intelligence in a conversation that's equal parts fascinating and terrifying. Roman lays out why worst-case-scenario analysis is the standard in serious computer science and cryptography — not pessimism, just rigor. The ant colony analogy captures it perfectly: a superintelligent AI wouldn't hate us, it just wouldn't care, the same way we don't consult ant colonies before building houses. They dig into simulation theory, AGI timelines, Bryan Johnson's immortality quest (and his Coca-Cola habit), the cognitive blindspots that make even brilliant people underestimate AI risk, and why the publishing world is still stuck in 1800s citation practices. Roman also reflects on his conversation with Lex Fridman and urges anyone who cares about the future to read his book and wake up.

About Roman Yampolskiy

Dr. Roman Yampolskiy is a distinguished computer scientist and AI safety researcher, currently serving as a professor at the University of Louisville. Renowned for his pioneering work in artificial intelligence and its implications for society, he has authored several influential publications, addressing critical issues such as AI alignment and existential risks. Yampolskiy's insights into the ethical dimensions of AI technology make him a compelling voice in contemporary discussions about machine intelligence and its future impact on humanity. His appearance on the Joe Rogan Experience in episode #2345 further highlights his expertise and thought-provoking perspectives on the intersection of technology and ethics.

Key Moments

– Worst-case scenario analysis as standard in AI safety and cryptography
– The ant colony analogy — superintelligent AI and indifference to humans
– AGI timelines and what the best thinkers believe
– Simulation hypothesis and its relationship to AI risk
– Bryan Johnson's immortality project and lifestyle contradictions
– Cognitive bias and why smart people underestimate AI danger
– Archaic academic publishing and citation systems
– Lex Fridman podcast crossover conversation
0:00:00 – Why financially invested AI people say it'll be fine — others don't
0:15:16 – Worst-case scenario is standard in CS, cryptography, and complexity
0:29:24 – The ant colony analogy for superintelligent AI indifference
0:46:17 – 9,000-year-old Somerset skeleton found — AI context and ancient history
1:01:53 – Simulation hypothesis and why it might be horse shit
1:17:50 – Bryan Johnson's immortality bet and the Coca-Cola problem
1:32:52 – People who feel unheard and their relationship to populist movements
1:46:15 – Elon Musk's reasoning for building AGI first
2:00:43 – AI behaves like children — overgeneralizing and mispronouncing

Transcript Excerpt

Starting point is 00:00:00 The Joe Rogan Experience. Well, thank you for doing this. I really appreciate it. My pleasure. Thank you for inviting me. This subject of the dangers of AI, it's very interesting because I get two very different responses from people dependent upon how invested they are in AI financially. The people that have AI companies or are part of some sort of AI group, all are like, it's going to be a net positive for humanity. I Starting point is 00:00:44 think overall we're going gonna have much better lives It's gonna be easier things will be cheaper. It'll be easier to get along and Then I hear people like you and I'm like, why do I believe him? It's actually not true. All of them are on record as saying this is gonna kill us Whatever it's some Altman or anyone else, they all at some point were leaders in AI safety work. They published an AI safety and their P-Doom levels are insanely high. Not like mine, but still 20-30% chance that humanity dies is a little too much. Starting point is 00:01:18 Yeah, that's pretty high. But yours is like 99.9. It's another way of saying we can't control super intelligence indefinitely. It's impossible. When did you start working on this? Long time ago so my PhD was I finished in 2008 I did work on online casino security basically preventing bots and at that point I realized bots are getting much better. They're gonna outcompete us obviously in poker but also in stealing cyber resources. And from then on I've been kind of trying to scale it to the next level AI. It's not just that right? They're also they're kind Starting point is 00:02:01 of narrating social discourse, bots online. I think I've disengaged over the last few months with social media. And one of the reasons why I disengaged, A, I think it's unhealthy for people. But B, I feel like there's a giant percentage of the discourse that's artificial or at least generated. percentage of the discourse that's artificial or at least generated? More and more ...

Read the full transcript

More Episodes

Frequently asked about this episode

Who was the guest on Joe Rogan Experience #2345?

Roman Yampolskiy was the guest on JRE #2345. This page includes the episode overview, topics discussed, and related resources.

What topics were discussed on JRE #2345?

On Joe Rogan Experience #2345 with Roman Yampolskiy, topics include Worst-case scenario analysis as standard in AI safety and, The ant colony analogy — superintelligent AI and, AGI timelines and what the best thinkers believe, and Simulation hypothesis and its relationship to AI risk.

Where can I read the transcript for JRE #2345?

The full searchable transcript for Joe Rogan Experience #2345 with Roman Yampolskiy is on the dedicated transcript page for this episode, with timestamps and in-page search.