@enigmatico Only sort of related, but kinda off-topic: I find it fascinating how people think LLMs gonna lead to AGI
If anything it actually prevent it from happening, because companies are now preoccupied with making "better" token predictors rather than try to further the machine learning research. (U know, the thing OpenAI non-profit was created to do, before it went full corpo mode)
Architecture upon which LLMs are built will never lead to AGI, its not meant for this, it could never do this.
And it sucks, not even cus of AGI, idc about that, but machine learning truly is a fascinating field of computer science, I was insanely into it as a child, but all this now just leaves dirty taste in my mouth. Because of course corporations had to ruin it