Gary Marcus endorses middle-ground AI risk framing after Hugging Face hack
GaryMarcus · x · 2026-09-16
Gary Marcus endorsed danprimack's take that the Hugging Face breach points to a plausible future of attacks on vital infrastructure — water, energy, food supply, or even weapons systems. "You don't need to wipe out humanity to hurt lots of humans"; this middle ground between doom and regulatory capture may be where consensus is possible. Marcus quipped that's probably why it got fewer than 50 likes.
More from Safety
- Bengio says AI regulation nears a Covid-style pivot as safety incidents pile up — nordicinst · 2026-09-16
- Australian online content could be scraped for AI by default under proposed copyright reforms — nordicinst · 2026-09-16
- Startups push back as Anthropic and OpenAI champion safety coordination that may squeeze smaller rivals — nmasc_ · 2026-09-16
- Senate committee advances JAWBONE Act to curb government pressure on platforms and AI — neil_chilson · 2026-09-16
- Synthesia exec: "AI is less regulated than a sandwich shop" claim is false — alexvoica · 2026-09-16
- US rushes medical AI agents for diagnosis and prescribing despite thin evidence — EricTopol · 2026-09-16