>>714474786I do lean a bit towards this. I work in a big company as IT Administrator. I see that there are attempts of incorporating AI at every step, everything related to chatbots is nearly useless. Sure, there are things that are good like image recognition to automatize some work on production lines or looking for error in end product(this can replace some of the low wage workers) but so far, every chatbot is a failure. First of all, for it to be usable in real way you have to have it on prem(locally) because if you want to feed it information from inside company it can't spill outside, which is costly even for big company, for the effect that it will have(faster searching of company data). Another problem is that, for specialized work like mine, even online chats with massive databases are useless. I tried to use it in many ways, it sometimes gives me wrong answers about thing that are available in open web documentation, not to mention it doesn't tell you it doesn't know, it spew false bullshit which defeats the point, I have to double check via internet anyway.
The only ever use case that I found in my work is when I want to make some easy powershell/python macro, it is faster to fix AI retardiness in easy tools than to make them from ground up and that's it.
All the channels on YT think that AGI is coming are retards who drink propaganda, there is already lack of data for further learning of the AI, hence absolute dystopian peaople from OpenAI want to make "wearable" to gather data from real world. Imagine, none of your conversations really private.