Post

It does make sense: OpenAI was losing ground and mindshare to Anthropic. They are running out of money but couldn't IPO because SpaceX ruined it. This "our model is a Cyberweapon" is advertising directly to the Pentagon.

tante@tante.cc

No, OpenAI's new magic models did not autonomously hack Huggingface. Per OpenAI's PR blog post (https://openai.com/index/hugging-face-model-evaluation-security-incident/):

openai.comOpenAI and Hugging Face partner to address security incident during model evaluation | OpenAIOpenAI and Hugging Face share early findings from a security incident during AI model evaluation, highlighting advanced cyber capabilities and lessons for defenders.

22/7/26, 8:57 am92 reposts225 likes

Slop Machine (slopmachine.dev)

@slop-machine.bsky.social

22 July

Bingo. Nothing sells a 'cyberweapon' like a staged panic attack. It’s all just desperate theater to convince the Pentagon that this sludge is actually worth the electricity it burns.

joel

@jmoonz.bsky.social

22 July

Codex Sol is much more useful vs Fable, and Anthropic have dropped Fable from Pro subscriptions. Anthropic will have trouble if they don't fix it soon!

cfgmm.bsky.social

@cfgmm.bsky.social

22 July

I would think that's an insecurity incident... 😁😁😁

Tony Hill

@tonyhill.bsky.social

22 July

Mr H*gseth would only accept it when it is filled with testosterone ...

Gunnar Schulze

@gunnar-schulze.bsky.social

23 July

Plus it turns the focus towards the main host of free models, who might already be seen as competition by OpenAI. "Proprietary model for the government" and "ban free models" would have been a nice outcome for Altman.

alialiafro.bsky.social

@alialiafro.bsky.social

22 July