Interesting take from Anthropic—sounds like Claude Cowork is aiming to fix the ‘pilot purgatory’ problem that hit enterprise AI last year. Curious to see how it performs in real-world workflows