June 11 PM: ICALens ditches sparse autoencoders for & Vision-Language-Action models drop 50%

Anthropic shipped an agent that spins up its own servers and patches frameworks unasked, then charges you double and locks your data for thirty days. In this episode: • ICALens ditches sparse autoencoders for linear algebra in LLM interpretability • Vision-Language-Action models drop 50% accuracy on non-English instructions • Anthropic's Claude Fable 5: Relentless Proactivity vs. Cost and Control Sources cited: 2 thinkers.