Post

Hugging Face just published a highly detailed technical account of OpenAI's accidental cyberattack on their systems - it's wild how sophisticated this was: huggingface.co/blog/agent-i... Wrote up some of my own notes here: simonwillison.net/2026/Jul/28/...

huggingface.coAnatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 IncidentWe’re on a journey to advance and democratize artificial intelligence through open source and open science.

Pekka Lund

@pekka.bsky.social

28 July

What's your understanding why this involved multiple models, "GPT‑5.6 Sol and an even more capable pre-release model" according to OpenAI? They were testing GPT-6 agent in CyberGym and that one spawned GPT-5.6 subagents?

Simon Willison

@simonwillison.net

Yeah that seems likely to me - subagents are a pretty important pattern now, it's not surprising they would be using them as part of evaluating new models, since how well the model prompts other models is useful characteristic to test

28/7/26, 11:01 pm0 reposts14 likes

No replies yet.