Matthew Green: assume models keep improving at the same rate as since early this year
matthew_d_green · x · 2026-10-12
Cryptographer Matthew Green offers his baseline for reasoning about AI capabilities: assume models will keep getting better at roughly the same rate they have since early this year.
More from AGI Musings
- Personal software will be huge — and tastemakers will be the new value — alexmacgregor__ · 2026-10-12
- "Commoditize your complement" dynamics are becoming exponentially more important in the AI trade — inductionheads · 2026-10-12
- Mathematician's real worry: AI companies are exploiting math's culture of openness — sokrypton · 2026-10-12
- Why is AI art gatekeeping creativity into an exclusive club? — VraserX · 2026-10-12
- 'If you're mid, you're extinct': the AI force-multiplier debate — Kyrannio · 2026-10-12
- Agent swarms: personal agents earning resource shares, 'fleets' rebrand predicted — seanwbren · 2026-10-12