Because that would take a minimum amount of effort, nuance, and reasoning, and the result would probably generate less interactions compared to a cheap shot based on vibes.

This isn’t surprising at all. It reminds me of staunch Apple haters who recycle superficial talking points as opposed to Apple nerds who have long lists of very pointed critiques.

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.

Clearly there are a plenty of people incorrectly blaming AI for bluesky outages, why indeed?