4 ms·
You trust LLM responses? I feel they are not worth the paper they are written on
by dylan604 3y ago
You trust LLM responses? I feel they are not worth the paper they are written on
- ripley12 3y agoYou don't necessarily need to trust them for them to be useful. I can often evaluate a Copilot autocompletion (check that code looks right at first glance, check that it compiles, hover over method+type signatures to see their docs, run the code) in less time than it would take me to find+read a Stack Overflow answer.
- capableweb 3y agoNo, no one sane would. Just like no one sane would trust random SO authors with any authority either. You still have to verify the information somehow, but at least in the LLM case I instantly get replies 24/7, and it's been reading more papers than any of the authors on SO.
- Legend2440 3y agoChatGPT has gotten quite a bit more accurate over the last few months. It does seem to have lost some creativity in the process, but it's much more true to the training data now. Code is easy to verify anyway.
- vbezhenar 3y agoI trust AI as much as I trust random answer on SO. It is a direction for further research and evaluation.
- rtpg 3y ago"ask LLM, then look up the language/library docs to see if the idea makes sense" has worked relatively well for me.