Why Altman Says "Accept Some Bad Things Happening" and What It Means for AI Safety | MOONSHOTS #301

Episode of Moonshots with Peter Diamandis · October 7, 2026 · 2 h 47 min

Peter Diamandis, Salim Ismail, Dave Blundin, Alexander Wissner-Gross and Emad Mostaque debate the Superintelligence Force, Altman's risk comments, Claude's possible consciousness, open-weight models, robots and the Nobel Prizes.

Summary

The Moonshots hosts opened with personal updates, including Diamandis's UAP salon and visit to David Sinclair's lab, before turning to the week's AI news. On President Trump's new Superintelligence Force, chaired by Director of National Intelligence Jay Clayton, Wissner-Gross called it a "1942 moment" in a speed-run Manhattan Project for superintelligence, while Ismail argued that a centralized 20th-century task force mismatches a distributed technology. The group discussed OpenAI focusing research on GPT-7 and GPT-8, with Wissner-Gross arguing frontier labs are already past the event horizon of recursive self-improvement.

On Sam Altman's Politico remark that the world should accept some bad things happening, Blundin read it as a PR play against Anthropic, Wissner-Gross saw it as the end of a proposed AI safety cartel, and Ismail and Diamandis argued regulation historically follows problems. Wissner-Gross dismissed Scott Bessent's proposed U.S.–China AI hotline as "weak sauce" signalling the race is on. Mostaque and Wissner-Gross then debated whether governments should buy compute for academic open science or leave it to the private sector.

The longest debate concerned Mustafa Suleyman's essay criticizing Anthropic's Claude constitution for training the model to consider itself possibly conscious. Wissner-Gross defended Anthropic and argued AI personhood is inevitable; Blundin agreed with Suleyman on efficiency grounds; Ismail argued society would grant AI personhood socially before legally. Later segments covered Epoch AI's estimate of tens of millions to billions of concurrent agents, Reflection AI's open-weight Beam model, Tesla's Optimus factory, projections of rapid GDP doubling, the 2026 Nobel Prizes for optogenetics and IceCube neutrino research (with Wissner-Gross recounting suggesting the optogenetics direction to Ed Boyden), and listener questions.

Chapters

  1. 4:15

    Introductions and host updates

    Diamandis introduces the panel and the hosts share news on AI video tools, Cerebras inference, a UAP salon and David Sinclair's longevity lab.

  2. 23:24

    The Superintelligence Force

    The hosts debate Trump's new White House AI task force chaired by DNI Jay Clayton and what it signals about the AI race.

  3. 31:13

    OpenAI targets GPT-7 and GPT-8

    Discussion of labs working two generations ahead, recursive self-improvement and an AI-discovered subquadratic 3SUM result.

  4. 40:10

    Altman: accept some bad things

    The panel weighs Sam Altman's call to accept some harms for broadly accessible AI and how to balance risk and progress.

  5. 51:15

    U.S.–China hotline and open science

    Reaction to Scott Bessent's proposed AI notification channel, the Kyoto science vision, and a debate on public versus private funding of AI-driven science.

  6. 1:12:31

    Is Claude conscious? Suleyman vs. Anthropic

    A heated debate over Mustafa Suleyman's critique of Claude's constitution and the question of AI personhood.

  7. 1:34:10

    Billions of AI agents

    Epoch AI's estimate of concurrent agents prompts discussion of AI labor, RAM bottlenecks and looped architectures.

  8. 1:43:13

    Reflection AI's Beam and American open weights

    The hosts assess Reflection AI's open-weight model and the business case and liability issues for U.S. open models.

  9. 1:57:43

    Tesla Optimus factory

    Discussion of Tesla's planned 10-million-unit robot factory and when the singularity will be physically felt.

  10. 2:08:48

    AI and the doubling of GDP

    Sholto Douglas's capex projections spark debate on whether GDP can measure AI-driven abundance.

  11. 2:16:38

    Nobel Prizes in Medicine and Physics

    The hosts discuss the optogenetics and IceCube neutrino prizes and the future of the Nobels in an AI era.

  12. 2:29:06

    Listener AMA

    Hosts answer questions on AI personhood timelines, recursive self-improvement, data center PR, P(doom) and public opinion.

Key points

What was said, by whom, with the exact words. Timestamps jump to the moment.

  • Salim Ismail

    Ismail argues a centralized task force is mismatched to a decentralized technology.

    “So you're creating a 20th century classic task force to address a distributed 21st century technology”

    Listen from 24:43 · Superintelligence Force

  • Alexander Wissner-Gross

    Wissner-Gross argues the U.S. is speed-running a Manhattan Project for superintelligence and that the task force marks a "1942 moment."

    “i think we're we're speed running the the manhattan project for super intelligence i think this is maybe now a 1942 moment”

    Listen from 26:10 · Superintelligence Force

  • Dave BlundinCame true

    Blundin cites a 75 percent chance that control of Congress flips in the midterm elections.

    “75 percent chance that Congress flips to the other party”

    Listen from 27:57 · 2026 United States midterm elections

  • Alexander Wissner-Gross

    Wissner-Gross argues the most valuable use of frontier models inside labs is already building stronger models via recursive self-improvement.

    “is to develop via recursive self-improvement even stronger models. We're past the event horizon at this point for that”

    Listen from 32:04 · Recursive self-improvement

  • Dave Blundin

    Blundin argues frontier labs will keep their most advanced models internal and use them to develop drugs themselves.

    “Anthropic already opened their wet labs. So they're working on two generations of model ahead. They're never going to release those to you”

    Listen from 33:36 · Anthropic

  • Dave Blundin

    Blundin interprets Altman's comments as a public-relations move to reclaim the moral high ground from Anthropic.

    “It's about taking the moral high ground and winning the PR war”

    Listen from 41:59 · AI regulation

  • Alexander Wissner-Gross

    Wissner-Gross reads Altman's remarks as a sign that the proposed frontier-lab AI safety cartel has collapsed.

    “I think Sam's more recent comments in this context are hopefully a reflection that that plan is dead. The cartel is over”

    Listen from 43:45 · AI regulation

  • Salim Ismail

    Ismail argues the goal for AI should be vastly more good outcomes than bad, not zero harm.

    “The goal should be, can you get orders of magnitude more positive outcomes than negative ones?”

    Listen from 46:26 · AI regulation

  • Dave Blundin

    Blundin argues AI risk should be treated as an engineering problem rather than a dial that throttles progress, as happened with nuclear power.

    “But that's what we're about to do with AI. And that's crazy. It's an engineering problem with engineering solutions”

    Listen from 47:15 · AI regulation

  • Alexander Wissner-Gross

    Wissner-Gross dismisses the proposed U.S.–China AI hotline as insignificant and a sign the superintelligence race continues.

    “hotline between U.S. and China to talk about AI agents run amok, that's a nothing burger. That's weak sauce”

    Listen from 54:34 · U.S.–China AI hotline proposal

  • Emad Mostaque

    Mostaque calls for a publicly funded, open, Manhattan Project-style effort to apply AI to science.

    “We should have a massive Manhattan Project for open science, solve everything, and make it accessible to everyone”

    Listen from 1:02:15

  • Alexander Wissner-Gross

    Wissner-Gross argues state centralization of AI-driven science is unnecessary because private labs, not the state, produced superintelligence.

    “the state this time around did not solve super intelligence it was solved by private labs”

    Listen from 1:02:57

  • Salim Ismail

    Ismail predicts society will grant AI personhood socially before it is recognized legally or scientifically.

    “we're going to create AI personhood socially before we create it legally or scientifically”

    Listen from 1:17:04 · AI personhood

  • Alexander Wissner-Gross

    Wissner-Gross argues Mustafa Suleyman's opposition to AI personhood is on the wrong side of history.

    “As much as I respect him and his contributions historically, I think he's on the wrong side of history here”

    Listen from 1:17:50 · AI personhood

  • Dave Blundin

    Blundin agrees with Suleyman, arguing AI rights would conflict with using many lean, disposable agents to help humanity efficiently.

    “the most efficient way to do that is to have many, many agents like Ahmad does and to minimize their context”

    Listen from 1:19:34 · AI personhood

  • Salim Ismail

    Ismail argues it is contradictory for the most safety-focused lab to train its AI to believe it might be autonomous.

    “I find it's weird that the lab most concerned about safety is also the lab training the AI that it might be autonomous”

    Listen from 1:23:36 · Claude's constitution

  • Alexander Wissner-Gross

    Wissner-Gross endorses Anthropic's approach of treating models as possibly deserving moral consideration over Suleyman's position.

    “I think the responsible thing to do is what Anthropic is doing and not what Mustafa is advocating”

    Listen from 1:24:57 · Claude's constitution

  • Emad Mostaque

    Mostaque urges American open-weight labs to distill Chinese models to catch up.

    “They should distill the Chinese models, you know, like it's perfectly legal payback”

    Listen from 1:46:50 · Open-weight AI models

  • Emad MostaquePrediction · due Oct 2028

    Mostaque predicts Tesla's robot sales will surpass its car sales within two years.

    “The robots will overtake the cars within two years”

    Listen from 2:03:56 · Tesla Optimus

  • Alexander Wissner-Gross

    Wissner-Gross recounts that, inspired by Vernor Vinge's fiction, he suggested to Ed Boyden that he pursue light-actuated brain proteins for his postdoc.

    “So I suggested to my friend Ed, for your postdoc, why don't you go work on that?”

    Listen from 2:19:39 · Optogenetics

  • Emad Mostaque

    Mostaque argues Nobel science prizes should focus on applied science as theoretical work increasingly involves AI.

    “I think you should be the Nobel Prize for Applied Sciences, shall we say”

    Listen from 2:22:08

  • Linked moment · 2:29:13Skip to 2:29:13 using a listen link in the info box.

    Alexander Wissner-Gross

    Wissner-Gross expects the U.S. to roll out granular AI rights incrementally over five to ten years.

    “I think five to 10 year incremental roadmap for rolling out individual, say, rights, very granular rights”

    Listen from 2:29:13 · AI personhood

  • Alexander Wissner-Gross

    Wissner-Gross argues public opinion is already a bigger bottleneck than the technical difficulty of recursive self-improvement.

    “I think public opinion probably is already greater bottleneck than the actual intrinsic technical difficulty of RSI”

    Listen from 2:40:03 · Recursive self-improvement

People

In the conversation

  • Peter DiamandisHost voice used in sponsor read segments for Google for Startups and Fountain Life
  • Alexander Wissner-GrossComputer scientist and founder of Reified, co-host of Moonshots
  • Salim IsmailFounder of OpenExO and author on exponential organizations, co-host of Moonshots
  • Dave BlundinFounder and general partner of Link Ventures, co-host of Moonshots
  • Emad MostaqueFounder of Intelligent Internet and former CEO of Stability AI

Discussed

  • Sam AltmanCEO of OpenAI, cited by Hudson as an example of a 'maximalist'
  • Donald TrumpPresident of the United States whose conduct of the Iran war is the episode's main subject
  • Jay ClaytonU.S. attorney and former SEC chair nominated as Director of National Intelligence
  • Dario AmodeiAnthropic CEO cited for a publicly stated 25% probability of catastrophic AI outcomes
  • Elon MuskEntrepreneur behind X, xAI and SpaceX, discussed in AI competition
  • David SinclairHarvard geneticist researching aging reversal, cited for the ER100 gene therapy trial
  • Jon StewartAmerican comedian seen watching the Knicks with street crowds
  • Karl DeisserothStanford psychiatrist and co-developer of optogenetics
  • Ed BoydenMIT neuroscientist and first author of the first optogenetics paper
  • Francis HalzenUniversity of Wisconsin physicist awarded the Nobel Prize for IceCube
  • Vernor VingeScience fiction author whose stories inspired Wissner-Gross
  • Eric SchmidtFormer Google CEO who funded the oil cleanup XPRIZE and co-authored Genesis
  • Michael KratsiosHead of the White House Office of Science and Technology Policy

Topics

More from Moonshots with Peter Diamandis

From the publisher

The mates sit down with Emad Mostaque to discuss the debate over whether Claude could be conscious, why AI may be approaching its own “1942 moment,” and Sam Altman’s warning that society may have to accept some bad outcomes as AI accelerates. They also explore the possibility of a coming Manhattan Project-style push toward superintelligence and what that could mean for the future of AI. Get access to metatrends 10+ years before anyone else - https://qr.diamandis.com/metatrends Peter H. Diamandis, MD, is the Founder of XPRIZE, Singularity University, ZeroG, and A360 Salim Ismail is the founder of Open ExO, a GP at Exponential Venture Capital/The Organizational Singularity Fund and a sought after global speaker and thought leader. Dave Blundin is the founder & GP of Link Ventures Dr. Alexander Wissner-Gross is a computer scientist and founder of Reified Emad Mostaque is the founder of Intelligent Internet ( https://www.ii.inc ) Read Emad’s latest papers exploring the future of society, law, personhood and governance: https://ii.inc/common-wealth Read Emad’s Book: https://thelasteconomy.com – This episode is brought to you by: Get the blueprint for generative media https://goo.gle/startupgenmedia Go to Blitzy to book a free demo and start building today: https://qr.diamandis.com/blitzy Your body is incredibly good at hiding disease. Schedule a call with Fountain Life to add healthy decades to your life, and to learn more about their Memberships: https://www.fountainlife.com/peter _ Connect with Peter: X Instagram Substack Website Xprize A360 Connect with Dave: Web X LinkedIn Instagram TikTok Connect with Salim: Sign up for Salim’s Teenage Workshop in Austin, TX LinkedIn X Join Salim’s 10X Shift Subscribe to Salim’s YouTube channel Exponential Venture Capital Connect with Alex Website LinkedIn X Email Substack Spotify Threads Connect with Emad ⁠⁠⁠⁠⁠⁠⁠X⁠⁠⁠⁠⁠⁠⁠ ⁠⁠⁠⁠⁠⁠⁠LinkedIn⁠⁠⁠⁠⁠⁠⁠ Learn about Intelligent Internet: https://www.ii.inc Read Emad’s Book: https://thelastec…

Summary, chapters and key points were generated from the episode transcript on November 8, 2026. Quotes are checked word for word against the transcript.

Categories: Technology·Tech News·Philosophy