[flagged]
What annoys me the most bsky AI hate is the assumption that people who spend a lot of time working with LLMs don’t understand their weaknesses, as if we aren’t constructing systems and evaluations to determine precisely how much AI sucks for our given task.