Beff Jezos: Private Post-Trained Models Near Frontier, and the 'God Model' Camp Wants Red Tape to Kill Them
beffjezos · x · 2026-09-13
e/acc founder Beff Jezos (Guillaume Verdon), quoting a post noting open-source models now match frontier closed models at 1/10th the cost just as calls to "hit the brakes" emerge, says private experiments show post-training base models gets near frontier on specialized tasks. He argues an "economy of individuated intelligence" threatens the centralized "God model" approach — and that its proponents want to shut it down with red tape.
More from AGI Musings
- Lab insider explains the internal-external gap on AI progress and Astra scare — MajmudarAdam · 2026-09-13
- Hugging Face's Noyant: Embodied AI saves lives but isn't profitable under capitalism — mervenoyann · 2026-09-13
- METR brings in Redwood staff for independent investigation of Anthropic alignment incidents — Jsevillamol · 2026-09-13
- MIT publishes official AI and Education report with eight guiding principles — alex_verem · 2026-09-13
- MIT report: 46% of undergrads use LLMs daily, 90% worry about overreliance — alex_verem · 2026-09-13
- "Doomsday sells tickets": skeptic calls Anthropic testimony a hype playbook — gerardsans · 2026-09-13