Looking to hear your guys’ thoughts on this, and hopefully share points in a more sophisticated manner than I can describe. (also, I hope this is an appropriate place to post?)
I have ran into this discussion a few times across the fediverse, but I can’t for the life of my find those threads and comments lol
I believe that a non-corporate owned platform with user-generated information is most optimal, like wikipedia. I don’t know the technicalities, but I feel like AI can’t replace answers from human experiences - humans who are enthusiasts and care about helping each other and not making money
I don’t know much about this topic, but I’m curious if you guys have actual real answers! Thread-based services like this and stack overflow (?) vs chatgpt vs bing vs google, etc.
Machine learning seems to be very good at generating believable persuasive writing, and not at all good at determining truth from fiction, even worse than people. This is an absolutely deadly combination and our rush to use it in this capacity is profoundly stupid.
I’m not against these algorithms mind you. I think they have a lot of useful potential. It’s just that the first things people have dived for to use it seem to me to be the absolute most foolhardy ways to apply it.
I completely agree. It makes sense that AI is not good at determining truth vs fiction. I think it’s more important for us as users to just search for information on our own, then determine the “end answer” with our own judgement after reviewing different sources and experiences (taking each individual answer with a grain of salt)
That’s why, I personally think AI search engine won’t be the best all-rounder for all types of information that’s not niche, deep searching which is IMO better found on forum-like platforms where people (enthusiasts) share sources, their experiences, what worked, what didn’t work, and why. For AI, maybe just simple bland information, like an excel formula, or how to hot wire a car, is better
yeah, AI does perform very well when given a specific and goal-oriented task. I think the coolest use I’ve seen for it was an emergency doc who was getting it to write explanatory documents for patients. Like “Please write a friendly, empathetic, simple-english explanation for why CPR would not be effective on a frail person with severe osteoporosis and advanced dementia” and things. This allows the doc to give the patient more detail than they’d have time to present, but it can be very closely tailored to the scenario, and it’s the sort of information AI shines at producing.