We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
The four strongest picks form a coherent stack: a long-form rebuttal to extinction claims, a policy map, an antitrust argument, and a data-provenance check. Start with The Contagion of Fear, which Martin Casado called “Exceptional. Please read.”
Start with The Contagion of Fear
- Type / creator: Blog post by @bcantrill.
- Recommended by:Martin Casado, with the unqualified endorsement above.
- Key takeaway: Using a confession about a fake computer-virus warning, the author argues that fear can take root and outrun correction. The rebuttal then asks readers to demand evidence proportional to extraordinary extinction claims and argues, from an engineering perspective, that intelligence alone is not enough: physical-world action, human-designed systems, accountability, and control still matter. The extinction claims being rebutted are explicitly attributed to other people, so this is a skeptical argument—not an independent risk estimate.
- Why it matters: It supplies a reusable test for AI-risk rhetoric: follow the proposed mechanism into the physical world and ask whether the evidence matches the claim, rather than treating a growing chorus of experts as evidence by itself.
Gavin Baker’s AI-policy summary
- Type / creator: X post by @GavinSBaker, opening with “Wild 24 hours for AI and lots of different proposals have been made.”
- Recommended by:Brad Gerstner, who called it a “Good summary” and a step toward avoiding reactionary over-regulation while balancing competition, open source, U.S. leadership over China, and public confidence.
- Key takeaway: Baker calls embedded third-party evaluators the only tangible new fact, then separates Dario Amodei’s broader package of national and international regulation from Sam Altman’s support for evaluators and Elon Musk’s proposal for competitor peer review before release. He closes on the distribution-versus-centralization question, arguing for broadly distributed AI rather than leaving intelligence with a few corporations.
- Why it matters: This is a compact map of the policy fault line: third-party evaluation is the shared operational step, while national/global regulation, antitrust waivers, open weights, and concentration of control remain contested.
BedoyaUSA’s AI-and-antitrust post
- Type / creator: X post by @BedoyaUSA.
- Recommended by:Chamath Palihapitiya, who said, “This is worth reading.”
- Key takeaway: The post argues that AI will not eradicate humanity; that companies can coordinate on safety without violating antitrust law, pointing to a DOJ/FTC policy statement on cybersecurity information sharing; and that antitrust should still prevent incumbent AI companies from coordinating to exclude cheaper rivals.
- Why it matters: It is a useful adversarial companion to Baker’s summary because it draws a sharp line between safety cooperation and market protection. Its sweeping claims are arguments to test, not settled findings, so read it for the distinction it makes.
Foreign accounts can drive the conversation on American AI policy
- Type / creator: X analysis by Future Jurvetson.
- Recommended by:Brian Armstrong, who shared it with the warning, “Beware the propaganda.”
- Key takeaway: Jurvetson’s post says Jacob Coxon’s Anthropic resignation post received 165 million views and that 76% of its reposts came from foreign countries, led by India and Indonesia. It contrasts that with Parker Thayer’s skeptical analysis, which received 6.6 million views and drew 34% of its engagement from foreign countries, then warns readers not to confuse what is trending with what Americans want in AI policy.
- Why it matters: This is a useful provenance check for the AI-regulation debate. The figures and interpretation are presented by the post itself, so treat it as a lead for investigating how political narratives spread—not as proof that foreign amplification caused a particular policy outcome.
The cartoon, sourced from waitbutwhy.com, does not feature a formal title but illustrates a graph plotting 'AI Intelligence' against 'Time.' The central subject is the exponential growth of artificial intelligence relative to biological intelligence. The central argument is that AI intelligence will rapidly surpass human capability—moving from 'monkey tricks' that amuse humans to a level of 'Reality' that is fundamentally incomprehensible ('The fuck??') to humans. This visual provides a succinct, conceptual explanation of the 'singularity' concept often cited by figures like Elon Musk, potentially serving as a non-promotional, illustrative resource for understanding AI acceleration risks.
Direct answer: The piece argues that AI-extinction fear is being irresponsibly seeded by technologists, amplified by public trust in experts, and advanced without evidence proportionate to its extraordinary claims; its counterargument is that engineering and physical-world control impose constraints that extinction narratives gloss over.
- Author/date status: The supplied article text contains no named byline or calendar publication date. It is written in the first person and refers to “my 2023 Monktoberfest talk,” but that reference does not identify the writer.
- Opening frame: The author recounts a university prank falsely warning humanities students that a computer virus was spreading. The resulting panic, shutdowns, lost work, and inability to retract the warning illustrate the article’s thesis that fear takes root and spreads faster than correction.
- Claim being rebutted: The post attributes to ex-Anthropic employee Jacob Coxon—and says Anthropic Alignment Science lead Evan Hubinger agreed—a claim that AI has a greater-than-10% chance of killing all humans within the next decade. It says proposed mechanisms such as hacking critical infrastructure and creating extinction-level bioweapons are hand-wavy and offered without further elaboration.
- Mechanism of fear: Because the public cannot understand every technical domain, it must rely on domain experts; the author argues that when experts raise alarms, fear propagates more readily than evidence, while the number of frightened experts can itself appear to confirm a consensus.
- Core rebuttal: From what the author explicitly calls a “narrow” expertise in building computers, the post argues that engineering requires physical-world action as well as intelligence. It says robots do not perform the feared actions today or on a timeline consistent with these extinction claims, and that digital systems’ physical agency remains mediated by human-engineered systems and human accountability and control.
- Useful recommendation takeaway: Present the post as a forceful, opinionated call for technologists to exercise maximal caution when raising alarms: extraordinary extinction claims require extraordinary evidence, and the burden of explaining LLMs, critical infrastructure, bioweapons, and extinction biology lies with those making the claims. The attributed risk claims and mechanisms should remain explicitly attributed, since the article’s discussion does not independently substantiate them.
- Future Jurvetson’s X analysis on foreign influence in U.S. AI policy — Brian Armstrong shared the post with the warning “Beware the propaganda,” directing readers to the analysis. https://x.com/FutureJurvetson/status/2098481615592951907 The post examines how foreign accounts amplify U.S. AI-policy debates and points to DigitalBorders.com for further analysis of foreign influence.
AI and antitrust X post — @BedoyaUSA: Chamath Palihapitiya shared this post with the explicit endorsement “This is worth reading.” https://x.com/BedoyaUSA/status/2099144876939948254 The post argues that AI is unlikely to eradicate humanity, that AI companies can coordinate on safety without violating antitrust law, and that antitrust should not shield incumbent companies from cheaper entrants; it also cites a DOJ/FTC policy statement on cybersecurity information sharing.
- Gavin Baker — AI-policy summary (X post): Brad Gerstner endorses it as a “Good summary” and says its approach is a step toward avoiding reactionary over-regulation while balancing competition, open source, U.S. leadership over China, and public confidence in AI’s benefits. Read the post The post synthesizes proposals for embedded third-party evaluators, competitor peer review, and a balance between broadly distributed and centralized AI capabilities.
Elon Musk endorsed an unspecified cartoon from Wait But Why, saying it “hits the 🎯.”
@bcantrill recommends @monktoberfest’s 2023 “anti-Doomer” talk as a current, accessible response to concerns about AI-driven extinction; the YouTube video is available here: https://www.youtube.com/watch?v=bQfJi7rjuEk
- The contagion of fear (blog post) — Martin Casado gave the piece an explicit endorsement: “Exceptional. Please read.” The linked resource is https://bcantrill.dtrace.org/2026/09/13/the-contagion-of-fear/.