Open-weight models now within four months of closed frontier models, analysis finds
pstAsiatech · x · 2026-09-20
A deep-dive analysis by Chamath Palihapitiya's team argues open-weight models have closed to within roughly four months of the best publicly evaluated closed frontier models, making the open vs. closed race a pressing commercial question.
Key points:
- Openness is a spectrum: fully open models, open-weight models with restrictions, and closed models with private weights.
- Palantir CEO Alex Karp warns enterprises risk surrendering their 'alpha'—proprietary knowledge and processes—to frontier labs.
- Satya Nadella makes a similar point: 'You essentially pay for intelligence twice, once with money, and again with the proprietary knowledge you must reveal.'
- These strategic-dependence concerns explain why more enterprises are exploring open models for control over data, infrastructure, and customization.
- The piece asks: if open models approach frontier performance, what are businesses still paying frontier labs for, and where does value accrue?
Related event: Report: Open-Weight Models Now Just 4 Months Behind Closed Frontier(4 posts)→
More from AGI Musings
- DeepMind releases AlphaGenome 2, predicting DNA variant effects across hundreds of cell types — osanseviero · 2026-11-12
- Ethan Mollick: 'talk to your documents' RAG solutions already feel dated — emollick · 2026-11-12
- Anthropic's pitch hasn't changed in 5 years: 'AI is winner-take-all, we will build god' — menhguin · 2026-11-12
- Peter Diamandis: We have 6-18 months to reach 'Safe Superintelligence' — what is Ilya building? — PeterDiamandis · 2026-11-12
- AI observer: headcount is no longer a proxy for ambition — signulll · 2026-11-12
- AI system proves longstanding quantum computing conjecture in three days, Lean-verified — _arohan_ · 2026-11-12