4 ms·
My issue with using LLMs for this use case is that they can be wrong, and when they are, I'm doing the research myself anyway.
by nunez 10mo ago
My issue with using LLMs for this use case is that they can be wrong, and when they are, I'm doing the research myself anyway.
- ragequittah 10mo agoThe times it's wrong has become vanishingly small. At least for the things I use it for (mostly technical). Chatgpt with extended thinking and feeding it the docs url or a pdf or 3 to start you'll very rarely get an error. Especially when compared to google / stack exchange.