The Autoregressive Self

LLMs, non-duality, and the gap between tokens

I’ve been staring at page 64 of I Am That by Sri Nisargadatta Maharaj¹ while reading AI researcher Andrej Karpathy’s – one of the best AI teachers of our time – latest thoughts on what makes intelligence “animal-like” versus something else entirely. Two completely different domains – spiritual philosophy vs machine learning architecture. But put them together, and some…