The Trade-off Between Open Weights and Security Defense
soumitrashukla9 · x · 2026-07-19
Boaz Barak states he supports OpenAI releasing gpt oss and generally favors open weights models, but he doesn't believe an absolute one-sided approach is correct.
His core point is that giving defenders early access to stronger cyber models offers security advantages. The quote tweet thread also debates whether open weights erode the moats of frontier companies, and whether the real issue is the model architecture itself or the business and safety narratives surrounding it.
More from AGI Musings
- WSJ: for Anthropic investors, a 10% extinction chance may still be a 90% portfolio win — peterwildeford · 2026-10-27
- Simile, the AI human-simulation startup, raises $300M+ at $2B valuation with 85%+ behavior prediction — joon_s_pk · 2026-10-27
- AI images are moldy bread, AI text is spoiled milk: a researcher's ick analogy — _vaishnavh · 2026-10-27
- AI images are moldy bread, AI text is spoiled milk, researcher argues — _vaishnavh · 2026-10-27
- Peter Diamandis: revisit the projects you abandoned because you couldn't afford the team — PeterDiamandis · 2026-10-27
- Nat Lambert: don't trust any AI model until you've rigorously tested it — natolambert · 2026-10-27