My timeline is flooded with the latest salvo of LLM shilling, with a bunch of posts that look like, "actually the agents are pretty good now, the people who said that AI doesn't work are obviously wrong about everything."

This is pushing my buttons because:
1. "Does this stuff actually work" was never the most important question, the shills are straw-manning all critiques of this tech
2. That being said, I remain convinced that LLMs are inherently flawed tools for programming computers