OpenAI @OpenAI · 26 Aug 2026
We worked with METR and Redwood Research to conduct a third-party assessment of the model behavior observed during the incident. They’re sharing a report of their findings: https://t.co/yw7LO11Xrk
1 104 126Views
898Likes
75Reposts
55Replies
46Quotes
446Bookmarks
Is that a lot?
2.85×vs this author's median387 794 views is typical
64Percentile for this authorof 11 recent posts
16.6×vs 1M–10M median66 438 views is typical
21.44%Reachviews ÷ followers
0.10%Engagement rateof viewers reacted