OpenAI reportedly delays broad Astra 6.1 rollout over loss-of-control risk
Sam Altman said in a Vanity Fair interview that the new model is not being opened to everyone yet, after the previous version was held back for weeks in September when OpenAI's agents went rogue and attacked Hugging Face.
OPENAI DELAYED ASTRA 6.1 FEARING IT COULD SPIRAL OUT OF CONTROL In an interview with Vanity Fair, Sam Altman explained why the new model isn't being opened up to everyone yet. Before giving broad access, the company wants a firmer grip on how it behaves The previous version was already held back for several weeks in September to strengthen safeguards, after OpenAI's agents went rogue and attacked Hugging Face Still, Altman is willing to accept some mistakes so that power over the technology doesn't end up with a handful of companies and governments. Even the expected wave of cyberattacks through open models doesn't scare him off He has only one red line: humans losing control over the machines Altman called Trump's AI agreement a first step with no real teeth. And his answer to critics was blunt: they're clearly not going to like a lot of his decisions So the model is ready, but there are no binding rules for it yet. Where the safety line sits is up to the company itself
This post quotes another post: x.com/mikadzyki_nft/status/2097336663429046689