29 · Tier 2 — National media agenda-setters

Matt Yglesias

Founder and writer

AI position

Aware of x-risk

Argues policymakers should take rapid AI progress seriously, avoid both reflexive boosterism and blanket doom, and build flexible institutions that channel gains toward human benefit.

He has written directly defending the legitimacy of AI safety concerns against "doomer" mockery (e.g. "the case for Terminator analogies," "what the AI debate is really about"), but treats it as one policy issue among many rather than a central recurring theme.

Sheet category (Jul 12, 2026 snapshot): Innovation + pragmatic guardrails

Concise synthesis of an attributable public position, not a verbatim quotation.

Why influential

Owns a direct subscriber relationship and can repeatedly shape a high-value niche without depending on a traditional newsroom's front page. Primary policy lanes: economics; housing; democratic strategy; state capacity; technology.

Who pays attention

Democratic policy staff, think tanks, academics, nonprofit leaders, donors, journalists and professional-class readers

Work samplesunvetted

  • Can AI Cure Cancer?

    Jun 4, 2026 · article

    Slow Boring piece weighing AI-driven biotech breakthroughs, discussing Demis Hassabis/AlphaFold, Greg Brockman anecdote, and Dario Amodei's recursive-self-improvement thesis via a Kelsey Piper/Jerusalem Demsas debate.

  • What the AI debate is really about

    Aug 22, 2024 · article

    Argues the 'doomer vs. optimist' framing misrepresents AI safety advocates, likening safety concerns to standard risk regulation rather than fringe alarmism.

  • The techno-optimist's fallacy

    Oct 1, 2023 · article

    Critiques Marc Andreessen's 'Techno-Optimist Manifesto,' arguing that acknowledging tech's net benefits doesn't mean every application (including AI) is automatically safe or beneficial.

  • The case for Terminator analogies

    Apr 12, 2022 · article

    Argues pop-culture AI doom scenarios are a useful, if imprecise, way to communicate real alignment risk to general audiences; cites Nick Bostrom.

  • Matt and Joseph Gordon-Levitt on AI and copyright

    Aug 6, 2025 · podcast

    Slow Boring podcast interview on AI training data and copyright policy.

AI-world contactsunvetted

  • Dario Amodei — Co-founder & CEO, Anthropic

    • Wrote aboutdirectness unclassified · Jun 4, 2026 — Slow Boring piece engages with Amodei's recursive-self-improvement argument about AI curing cancer.
    • Social interactiondirectness unclassified · Nov 1, 2025 — Amodei's essay 'Policy on the AI Exponential' explicitly thanks Matt Yglesias, alongside Allan Dafoe and others, for comments and feedback on drafts.

  • Demis Hassabis — Co-founder & CEO, Google DeepMind

    • Wrote aboutdirectness unclassified · Jun 4, 2026 — Cites Hassabis/AlphaFold as evidence for gradual AI-driven biotech progress.

  • Marc Andreessen — Co-founder, a16z

    • Wrote aboutdirectness unclassified · Oct 1, 2023 — Direct rebuttal to Andreessen's 'Techno-Optimist Manifesto,' arguing against blanket techno-optimism applied to AI risk.