the best part about this new product is that you can interrupt, correct and clarify during long running tasks. it’s probably going to get overlooked in the discourse but it’s a major ux pain point and hopefully it sets a new standard for LLM interactions.