e/acc's Beff Jezos Argues Capability Diffusion Is Safest, Slams Anthropic's Closed Approach

beffjezos · x · 2026-09-15

Beff Jezos, quoting a DeepSeek engineer's essay, argues that rapid diffusion of AI capabilities is actually the safest strategy: all systems can harden as soon as a new capability ships. He claims holding models back behind closed doors and compounding advantages is what's dangerous — calling out Anthropic directly. The quoted essay highlights the OpenAI vs Anthropic divide: OpenAI opens its strongest models within weeks/months even at negative business EV.

Original post →

More from AGI Musings

AGI Musings channel →