3 ms·
We finally found a category of use cases for LLMs that have strong user acceptance: wading through endless bureaucracy. Even the impact of hallucination is min
by sublinear 3y ago
We finally found a category of use cases for LLMs that have strong user acceptance: wading through endless bureaucracy.
Even the impact of hallucination is minimal because the humans do that all the time!
- dmd 3y agoWhat does this have to do with LLMs?
- azemetre 3y agoIME LLMs are decent at summarizing text, I can see how it can help summarize various processes that must be done. I am unsure how it summarizes truly novel things tho, I would assume it might hallucinate more but haven't experimented much.
- dylan604 3y agoI wouldn't be surprised to see an account that used an LLM to do this very thing for any concept
- the8472 3y agoTheir bureaucracy-navigators will be no match for our bureaucracy-generators.
- IIAOPSW 3y agoIs this a generator-adversary network?
- rqtwteye 3y agoThat's my prediction. I have fed some extremely tedious enterprise contracts into ChatGPT and got a nice summary and answers for my questions. It wasn't perfect but with some tuning I think this will work really well.