3 ms·
> We aim to issue an alert within 30 minutes after concerning activity is surfaced through our monitoring system. If the monitoring system identifies a likely v
by dkoy 2mo ago
> We aim to issue an alert within 30 minutes after concerning activity is surfaced through our monitoring system. If the monitoring system identifies a likely violation of a critical security boundary, it generates a highest-priority alert. In our current implementation, the safety, security, and research teams are paged. If they cannot conclusively determine within 30 minutes that the flag is a false positive, those teams are expected to pause the activity.
Can't a lot happen within ~60 minutes?
- georgemcbay 2mo ago> Can't a lot happen within ~60 minutes? 60 minutes is a long time for a human attacker to do damage. With an LLM attacker it is an eternity.
- reasonableklout 2mo agoThe HuggingFace breach took place over two-and-a-half days [1], so 60 minutes is certainly better than nothing. [1]: https://huggingface.co/blog/agent-intrusion-technical-timeline https://huggingface.co/blog/agent-intrusion-technical-timeli...
- chrisjj 2mo ago> Can't a lot happen within ~60 minutes? Spawn a ton of unpausable processes, I'd say.
- andai 1mo agoThey just made it ~10x faster with the Cerebras deal, so that's the equivalent of 600 minutes in pre-Cerebras time.