Comment by intended

LLMs are narrative machines, not analysis machines.

The article isn’t about code, and on HN we default to that all the time.