Why can't I pay OpenAI or Anthropic to hack my systems with unsafe models, asks David Holz

DavidSHolz · x · 2026-09-10

David Holz questions why no lab offers a paid service letting unsafe models penetration-test your systems, with safe models then advising fixes — pointing at a missing product between model safety tiers and legitimate enterprise red-teaming demand.

Original post →

More from Models

Models channel →