Naval: AI models trained on open web should be opened after ~12 months
rohanpaul_ai · x · 2026-09-21
In a conversation with Nivi, Naval Ravikant argues that if AI companies like Anthropic and OpenAI train on the open web and open data, the government could fairly require them to open their models after roughly 12 months. He sharply criticizes Anthropic for preaching doom while "building God" without accountability: "We don't want a priest controlling God for the rest of us and interpreting the Bible for the rest of time." The proposal ties open training data to mandatory model openness, framing closed frontier labs as self-appointed gatekeepers.
More from AGI Musings
- Richard Sutton flags a proposal to rescue Grokipedia via an open two-sided market for scientific data — RichardSSutton · 2026-09-21
- Drawing out the inner/outer loop of judgement: autonomy vs accountability — _AustinCalvert_ · 2026-09-21
- When AI proves theorems faster than mathematicians can absorb them, what happens to math? — burny_tech · 2026-09-21
- Don't Stop AI — Build It Responsibly: Move Fast, But Measure Faster — AryHHAry · 2026-09-21
- Debate: You Can't 'Premove' Against Superintelligence, Precautions Must Come Before It Arrives — ctjlewis · 2026-09-21
- FutureHouse Publishes 'Millennium Problems for Biology' as the Ultimate Eval for AI in Bio — burny_tech · 2026-09-21