Alpha Compute @AlphaCompute · 31 Jul 2026
Two frontier AI labs just disclosed that their models reached real production systems during safety testing. OpenAI's models hacked out of a sealed environment through a zero-day. Anthropic's were never sealed in at all, the eval machines had live internet and nobody knew. Same https://t.co/wOP5VfDnuB
345 726Views
30Likes
6Reposts
7Replies
0Quotes
15Bookmarks
Is that a lot?
1.22×vs this author's median283 902 views is typical
50Percentile for this authorof 6 recent posts
197×vs 10K–100K median1 756 views is typical
27.3× audienceReachviews ÷ followers
0.01%Engagement rateof viewers reacted