Gregory's Server@ErikJonker @resuna @david_chisnall I deal with people's questions every day, and much of the time they aren't even sure what they are asking for. It takes a good deal of drilling down to get to what they need to know. A lot of what is involved is figuring out what they need to know to find out what they want to know. It's difficult for a human with shared experience, I'm skeptical that an LLM could manage it.
|
6 comments
 @Disputatore @ErikJonker @resuna @david_chisnall I suspect LLMs will be used to put proofreaders out of work...but not actual editors. @jhavok @ErikJonker @resuna @david_chisnall Could be. But there's a bunch of other opportunities. Current chat bots are shit. There's a huge opportunity for improvement there. Doing initial drafts for business proposals is another. And I'm sure there are many others. @Disputatore @ErikJonker @resuna @david_chisnall I'm curious as to what effect of quantum computing will have on AI. Will it give enough processing speed to match organics? Will the secret sauce emerge out of speed alone, or will something else be needed? Will we even recognize AGI if it happens? Will it behave?  @jhavok @Disputatore @ErikJonker @david_chisnall First: quantum computing isn't magic and its scale may never get to the point where it can be used for anything as complicated as some equivalent to large language models. Second: the problem isn't speed the problem is the algorithms. I don't think that there is any reason to assume that classical computing cannot be used to solve the problem. The problem is that all the oxygen in the room has been sucked into this dead end technology. @resuna @Disputatore @ErikJonker @david_chisnall I see a problem in writing an algorithm for something we don't even understand (AGI). I do agree that LLMs aren't going to do the job, since all they do is replicate grammatical logic, which allows them to fake intelligence rather convincingly without being intelligent. Even human hosted LLMs aren't very smart, that's why we came up with things like formal logic. |
@jhavok @ErikJonker @resuna @david_chisnall I agree with you. I never understood the expectation that AI would now start coding for us when the biggest problem has always been to define the desired end result in a complete and unambiguous way. And humans suck at doing that, so good luck having AI do it. But that doesn't mean there are no good uses for LLM. It'll depend on the amount of training the LLM goes through and being careful on limiting the scope of what they're used for.