Import AI 457: AI Stuxnet, Cursed Muon Optimizer, and Positive Alignment
Import AI 457: AI stuxnet; cursed Muon optimizer; and positive alignment
Import AI 457 covers fast16, Aurora, and positive alignment: SentinelOne found fewer than 10 matching files for fast16 signatures, while Tilde Research reports Aurora reached 2.26 loss on 1.1B-parameter transformers versus Muon’s 2.31 under a ~100B-token setup.
Why it matters: SentinelOne's teardown of fast16.sys shows it tampering with high-precision results in memory five years before Stuxnet, while Tilde Research finds Muon kills over a quarter of MLP neurons early in training.