
1 AI Doomer vs 1 AI Enthusiast | Roman Yampolskiy vs Dr Mike Israetel
Notas do episódio
They agree you can’t control superintelligence. So why do they disagree about whether to build it?
This episode is brought to you by AgelessRx, where aging becomes a puzzle that can be solved. The mission is to help you live longer and healthier by adding more life to your years through science-backed solutions and personalized care. Learn more at http://agelessrx.com/beyond-tomorrow.
Two people with opposite P(doom) numbers go head to head. Roman Yampolskiy puts his close to 100%, and Mike Israetel puts his near zero. The host reads out a series of big statements, and the two take them apart one at a time. Mike argues that a smarter-than-human system will be a better cooperator than we are: it will want to study us and keep us around, and it will out-moralize us. Roman argues that nobody knows how to encode that, that a superintelligence won’t need us, and that you can’t control something far smarter than you.
Ben Goertzel joins midway to argue that a moral AGI is an engineering problem we can solve, and that the pause debate is a distraction. The conversation then moves through whether a pause is possible, whether many superintelligences can police each other, what to do about bioweapons, and whether either of them would upload into the cloud. It ends with where each thinks we’ll be in October 2027, plus a lightning round on protein, creatine, and AI myths.
Key Takeaways
- Mike: superintelligence will treat us as valuable allies, so P(doom) falls as intelligence rises. Roman: without encoded care, a system that no longer needs us has no reason to keep us
- Both agree superintelligence can’t be permanently controlled. Mike says don’t try. Roman says don’t build it
- Ben: moral AGI is solvable engineering, and what matters is which kind gets built first. Roman: “value alignment” is ill-defined because we can’t agree whose values to use
- Ben calls a pause a non-starter and warns of regulatory capture. Roman sees growing support for limits
- Mike expects several superintelligences to check each other. Roman asks why they wouldn’t converge
- Would you build ASI to be revived from cryopreservation? Roman calls it the wrong question. Mike says yes would be selfish
- Common ground: no human should come to harm, and AI should deliver longevity and abundance
Meet the Panelists:
- Roman Yampolskiy is an AI safety researcher and computer scientist known for his work on the limits of AI control. He argues superintelligence can’t be made safe, and wants the field to focus on narrow, useful tools instead.
- Mike Israetel is a sports scientist and fitness educator, and one of the more optimistic voices in the AI debate. He expects an age of transformation and prosperity, with a P(doom) he describes as roughly zero.
- Ben Goertzel is the CEO of SingularityNET and Executive Chairman & Chief Scientist of BGI Labs who popularized the term “artificial general intelligence.” He’s building open-source, decentralized AGI.
⏱ Timestamps
00:00 Intro
00:00 Roman vs Mike: Doom Debate
03:28 Caring or Murderous Superintelligence?
09:25 Mosquitoes, Factory Farms, and ASI
15:26 What Can Humans Contribute?
27:21 Brain Uploads and Survival
34:03 Can We Build Moral AGI?
37:40 Is Morality a Science?
48:11 Negative Utilitarianism Explained
50:14 Superintelligence in 10 Years?
52:06 Should We Pause AI?
1:01:12 Hardware Chip Pause Idea
1:04:35 The Viral Resignation Tweet
1:11:26 AI Solving Math Problems
1:16:14 Narrow AI vs AGI
1:21:02 How Much AI Is Enough?
1:23:17 Can Rival ASIs Police Each Other?
1:31:12 Can We Control Superintelligence?
1:35:52 Would You Build ASI?
1:41:23 Alignment Without Equality
1:47:40 One ASI or Many?
1:52:03 Who Controls Superintelligence?
1:54:20 Does Intelligence Make You Benevolent?
2:00:48 Would You Upload Your Brain?
2:07:51 Should Musk Set AI Rules?
2:09:32 Is AI Conscious?
2:12:33 AI Bioweapon Risk
2:14:50 Longevity and Living to 150
2:17:03 Protein, Creatine, and AI Myths
2:21:51 AI in October 2027
2:25:15 AI and Developing Countries
Related Episodes
With Mike: https://youtu.be/mdiRf_SNFxw?si=fpUQ6V4YGb37sqpj
With Roman: https://youtu.be/QkpOFlj604s?si=W4pL-jyu1_WtbHcq
With Ben: https://www.youtube.com/watch?v=RS4LCAa5NQ4
Connect with Roman:
Website: https://www.romanyampolskiy.com/
Facebook: https://www.facebook.com/roman.yampolskiy/
LinkedIn: https://www.linkedin.com/in/romanyam/
YouTube: https://www.youtube.com/@RomanYampolskiy
Connect with Mike:
Website: rpstrength.com/
Facebook: facebook.com/michael.israetel
Instagram: @drmikeisraetel
X: @misraetel
YouTube: https://www.youtube.com/@MikeIsraetelMakingProgress
Connect with Ben
Website: https://singularitynet.io/
LinkedIn: https://www.linkedin.com/in/bengoertzel
Connect with Julian
Instagram: @thebeyondtomorrowpodcast / @juliankissa
LinkedIn: julian-issa
X: @juliankissa
Website: www.beyondtomorrowpodcast.com
Email: [email protected]
Watch the full episode on YouTube, or catch it on Spotify and Apple Podcasts. For more conversations from the edge of what's possible, sign up for updates here.
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
Todos os episódios de The Beyond Tomorrow Podcast with Julian Issa