Comment by rv3392
1 day ago
ML is still a thing. I believe that most AI research is still non-LLM ML-related - things like CNN+Computer Vision, RL, etc. In my opinion, the hype around LLMs has a lot to do with its accessibility to the general public compared to existing ML techniques which are highly specialised.
To be fair, I remember that some 5 years ago a lot of ML was quite accessible to programmers as it was often just a couple lines of python using tensorflow, or later pytorch.
I am almost in disbelief that LLMs are the thing that reached the "tipping point" for most companies to magically care for ML. The amount of products, that could have been built properly 5 years ago, that exist now in a slower form because of "reasoning" LLMs, is likely astonishing.
scikit too, i think. LLMs probably took off because they think it's a legal shield against "IP infringement/theft". Amazon has access to whatever magnitude of books and so does google, so if FB, X, mistral (actually i am not sure, is that a university led project? they probably got books if so), openai want to return decent results about books, they gotta get the books, too. Buying, scanning, OCR, copyediting, feeding the training scripts (jsonifying the input most likely), forget that, anna and russia and the high seas are literally right there, 4 octets away.
I'd hope that the very first commercially successful "AI media" be it a 1 minute commercial or a 10 minute TV segment or whatever brings the lawsuits. I really want to know if i can feel any vindication about arguing about this for the last 3 decades of my life. (IP specifically)
more to your member-berries, whole swaths of interesting research disappeared, either abandoned or bought and closed sourced. Genetic Algorithms, artificial life, stuff with optics, 3-atom thick transistors (hey, IBM patented that, but microsoft also did basically the same thing with their STP qubits - everything has to be arranged at atomic widths or whatever. IBM also built a USS Enterprise (unsure if D i am not a huge fan) out of atoms, forget if to scale. in like 2003. Microsoft spent 17 years playing catch-up with *the* hardware people.)
yeah. is the conclusion that moneyed interests suck?