Voidgolem posted...
the benefit of the "AI" being sold is a productivity boost to the existing human workforce. This is questionable and in most cases has caused a productivity drop because you still need people to spot-check the work (which is often incorrect)
for the other stuff it's actually useful for it's been in use for awhile, I think (machine vision/machine learning models for analyzing things - stuff a model is actually good for when given proper data)
You have, on demand, a very powerful search index for whatever you want, whenever you want...but people keep wanting to use this to spit out stupid generative nonsense that will never be useful.
I do not believe there is an avenue from the models that keep being (over)sold and tuned to keep people engaged with them, and the theoretical general AI that all the techbros want.
I agree and Id add I find it deceptive when people point to ML use in disease research as a reason to support commercial LLMs. You can still support disease research and also acknowledge the potential negative societal impact of ChatGPT and the like.