Software Engineers: Do you honestly get anything useful out of LLMs?

For 6 months now I've been trying to make agentic coding work for me, using Pi and a handful 30-120B models (Qwens, Nemotrons, Leguna...etc). I'm not greedy either, I stick to decent quants, never quantize kv cache, and keep my sessions up to 90k max. But the results have ALWAYS been disappointing.

中文全文翻译正在补齐，可先查看下方英文原文。

For 6 months now I've been trying to make agentic coding work for me, using Pi and a handful 30-120B models (Qwens, Nemotrons, Leguna...etc). I'm not greedy either, I stick to decent quants, never quantize kv cache, and keep my sessions up to 90k max. But the results have ALWAYS been disappointing.