FULL STORY
GPT-2, Seven Years Later: From Fringe to Consensus
Dario Amodei recalls being scared by GPT-2's 2019 release as scaling laws went from pseudoscience to consensus, reigniting debate over OpenAI's staged release strategy.
2026-10-28 ~ 2026-10-28 · 2 episodes · 7 posts
Episode 1 · Dario Amodei recalls GPT-2 moment: skeptics now seek AI jobs (2026-10-28, 4 posts)
Anthropic CEO Dario Amodei recalled being "scared" by GPT-2 in 2019, predicting million-fold scaling would surpass humans; the 90% who called his fears pseudoscience are now applying to AI labs.
- Dario: Those who called his GPT-2 fears 'pseudoscience' in 2019 now all apply to AI labs — irinarish · 2026-10-28
- Dario Amodei: scaling GPT-2 a millionfold would beat humans at everything — haider1 · 2026-10-28
- Dario Amodei: seeing GPT-2, he knew scaling a million-fold would beat humans at everything — haider1 · 2026-10-28
- Dario: people who called GPT-2 fears pseudoscience now apply to AI labs — zacharynado · 2026-10-28
Episode 2 · Seven years on, GPT-2's staged release still sparks debate (2026-10-28, 3 posts)
Anthropic co-founder Jack Clark revisited OpenAI's staged release of GPT-2 five years on, as commentators debate whether it set a valuable norm of caution or was needless 'safetyism' given the model posed mainly spam-level risks.
- Debate: GPT-2 was the last chance to delay the trillion-dollar cluster race — abhiadesai · 2026-10-28
- GPT-2 staged release debate revisited: setting a caution norm vs overhyped PR — gleech · 2026-10-28
- Jack Clark revisits GPT-2's staged release five years on: it was about setting a norm of caution — gleech · 2026-10-28