Noam Brown: just 2.5 years ago labs showcased frontier LLMs on grade-school GSM8K
polynoamial · x · 2026-10-27
Noam Brown, OpenAI researcher and co-creator of Libratus/Pluribus, reflects that only 2.5 years ago AI labs were still highlighting their frontier LLMs' performance on the grade-school math benchmark GSM8K — and he can't wait to see where we'll be in another 2.5 years.
Related event: OpenAI's Noam Brown notes GSM8K math benchmark is already obsolete(2 posts)→
More from AGI Musings
- Richard Socher: Transformer's real breakthrough was killing recurrence, not attention — DeryaTR_ · 2026-10-27
- Reddit user finds Instagram reel's comment section entirely AI-generated — stupidunluckytrader · 2026-10-27
- John Carmack: no robot today can change batteries in a jack-o-lantern — ID_AA_Carmack · 2026-10-27
- Stripe exec: AI trains on words yet is terrible at writing — jeff_weinstein · 2026-10-27
- Thiel asked what matters most post-SI, pauses 30 seconds: "Human Relationships" — kleffew94 · 2026-10-27
- Claude Opus 5.5 claims ETH false, all-dimension Kakeya and Yang-Mills in one day — repligate · 2026-10-27