Ask HN: The notion "LLMs are sucking the oxygen out of AI research"
Any research into Neural Networks is beneficial overall. There's so much research spurred by Transformers that it makes me think the original AI research is still going on but with less exposure.
We've seen massive networks forced into very small footprints which is insane. There have also been behemoth efforts in training at scale which also contributes to research, development, and IT infrastructure.
I love ML. I've been doing it since I was a kid in 2006 wondering what the hell a picture of a saddle had to do with making my Markov-like chatbot talk. (I accidentally reinvented a Markov chain without knowing what it was. It was a very bad implementation though and didn't work too well.)
My point is that now that the AI hype seems to be finally dying down, what is your opinion on the state of AI research?
Breakthroughs were the reason we got this far; I don't think LLMs have sucked the life out of AI, I think they've invigorated it and caused more people to be curious like I was many years ago.