4 ms·
Give an LLM all the time you want, and they will still not get it right. In fact, they most likely will give worse and worse answers with time. That’s a big dif
by dieortin 2y ago
Give an LLM all the time you want, and they will still not get it right. In fact, they most likely will give worse and worse answers with time. That’s a big difference with a software developer.
- mypalmike 2y agoMy experience is very different. Often it (ChatGPT or Copilot, depending on what I'm trying to accomplish) gets things right the first time. When it doesn't, it's usually close enough that a bit of manual modification is all that's needed. Sometimes it's totally wrong, but I can usually point it in the right direction.
- TeamDman 2y agoI mean, with a nonzero temperature, the randomness will eventually produce every combination of tokens in the corpus, so with a sufficiently large "all the time you want" you can produce limitless correct answers