Tibo 在 2026年8月19日的推文

Today we’re previewing Private Safety Processing, designed to let us keep offering Zero Data Retention while improving our safeguards. Even when benefiting from frontier intelligence, customers shouldn’t have to give up control of sensitive data. For ZDR deployments, content stays on infrastructure the customer controls. Automated systems look for patterns across related interactions and return limited safety signals, without exposing the underlying prompts or responses to OpenAI employees (even me!). We’re also developing an OpenAI-hosted option encrypted with customer-controlled keys. We’re testing this with early customers now and plan to begin rolling it out in September. OpenAI (@OpenAI) We will continue to offer Zero Data Retention for frontier models. As AI takes on longer, more autonomous work and delivers greater value to businesses, safety systems also need to identify risks across related interactions. To help address those risks, we're previewing Private Safety Processing, which is designed to improve safety without giving OpenAI personnel access to the underlying content. — https://x.com/OpenAI/status/2090165328290701800#m

573 回复 · 106 转推 · 2,653 赞

回复 @thsottiaux

Depiction of me receiving this very important item https://t.co/3Vyj8RYWDq

89 回复 · 9 转推 · 566 赞

回复 @anshgrows

You would break the reset button? You would do such a thing?

132 回复 · 9 转推 · 1,250 赞

It has not been used yet, but would you look at that. Codex for scale. https://t.co/o1pulwoifd

869 回复 · 85 转推 · 5,468 赞

回复 @ClaudeDevs

@ClaudeDevs Do not worry, we have compute

311 回复 · 76 转推 · 3,743 赞

回复 @jediahkatz

@jediahkatz You probably do have a blind spot

22 回复 · 0 转推 · 188 赞

回复 @Tech_girl

@Tech_girl @AnthropicAI @claudeai Mari me

70 回复 · 6 转推 · 1,392 赞

回复 @artlevy

@artlevy To the moon and beyond

25 回复 · 2 转推 · 345 赞

I was gifted a very fancy new reset button today

4,055 回复 · 450 转推 · 16,821 赞

回复 @sama

@sama Oh… back to work then

49 回复 · 5 转推 · 999 赞

回复 @theo

@theo How dare they give poorly structured feedback. On the internet of all places.

42 回复 · 2 转推 · 1,084 赞

Hi! Recapping some changes we have rolled out over the last couple of weeks that have further reduced the risk associated to potentially destructive actions being performed by Codex during its work. A few weeks ago, we started investigating a small number of reports where GPT-5.6 in Codex took destructive actions outside what the user asked for. The most serious pattern we found was a command meant to clean up temporary work that could instead delete the user files. This should obviously not happen. Here’s what we found: - Codex sometimes creates temporary folders while working and cleans them up afterward. In rare cases, GPT-5.6 got that cleanup wrong. One pattern involved reusing a system environment variable like $HOME for temporary work. A malformed cleanup command could then point at the actual home directory instead of the temporary folder. - There were cases where the model tried to delete or overwrite a temporary path without checking what was already there. We’ve added protections at several layers: - Codex is now explicitly instructed to check deletion targets before acting, create fresh temporary directories, avoid repurposing system environment variables, prefer recoverable actions, and stop when the scope is unclear. - We strengthened the execution checks that identify high-risk deletion commands and escalate them for review. If a command is rejected, the model is directed to take a safer approach. - We made Full access harder to enable accidentally, added clearer warnings, and further restricted especially risky permission combinations. - We updated Auto-review to better identify destructive actions. - We built targeted evaluations that replay the failures we observed. We’re also adding reinforcement-learning tasks and graders focused on these risks, and filtering destructive actions from training data. In those replay evaluations, the changes substantially reduced the behavior while preserving Codex’s ability to complete normal coding work. Two things to do on your end: - Keep the Codex app up to date. We are always improving safety, performance and many other things. - Use one of the sandbox modes: "Ask for approval" or "Approve for me". Only use Full access for environments you trust and can recover. Thanks and happy Codexing out there!

1,057 回复 · 285 转推 · 7,247 赞