Voices
Zvi ends five straight posts on the Hugging Face hack and calls recurrent depth 'playing with fire'
AI #184, published September 3, closes out Mowshowitz's five-post run on the Hugging Face incident with an unusually blunt 'I am exhausted.' He dismisses the week's model wave, Mythos 5.1, Fable 5.1, Gemini 3.8 Flash, Muse Spark 1.3, GLM-5.3-Flash and Astra, in a line: 'none of them look to be moments, or to upend the game board.' His one real alarm is Astra's recurrent depth, which he calls playing with fire while conceding the interpretability damage is currently manageable. He also picks up Dean Ball's self-sovereign AI argument and, citing February 2026 data, notes 'AI use within a firm did not have much impact on employment within that same firm,' while cautioning that measures within-firm effects only.
↳ Follow the thread