But we do have to acknowledge that AI is very much turned into an all encompassing term of everything ML. It is getting harder and harder to read an article about something being done with "AI" and to know if it was a custom purpose built model to do a specific task or is it throwing data into an LLM and hoping for the best.
They are purposefully making it harder and harder to just say "No AI" by obfuscating this so we have to be very specific about what we are talking about.
For a while I made an effort to specify LLM or generative AI vs AI as a whole,
but I eventually became convinced that it was no longer valuable.
Currently AI is whatever OpenAI, Anthropic, Meta, NVidia, etc say it is,
and that is mostly hype and marketing.
Thus I have turned my language on its head,
specifying "ML" or "recommendation system" or whatever specific pre-GPT technology I mean,
and leave "AI" to the whims of the Sams and Darios of SV.
I expect the bubble to pop in the next 3-6 months, if not before the end of 2025,
taking with it any mention of "AI" in a serious or positive way.
Wow, you are an optimist. I do feel "it's close", but I wouldn't bet this close. But I wouldn't argue either, I don't know. Also, when it really pops, the consequences will be more disastrous than the bubble itself feels right now. It's literally hundreds of billions in circular investing. It's absurd.
But we do have to acknowledge that AI is very much turned into an all encompassing term of everything ML. It is getting harder and harder to read an article about something being done with "AI" and to know if it was a custom purpose built model to do a specific task or is it throwing data into an LLM and hoping for the best.
They are purposefully making it harder and harder to just say "No AI" by obfuscating this so we have to be very specific about what we are talking about.