We Are (becoming) Ordinary

Part of my interest in machine learning roots from my own existential crisis. Only in recent years do we see all the abilities that we previously thought of as “intelligence” can indeed be largely reduced to problems of statistics. The bitter lesson is not only “bitter” to me in the sense of research but also in the sense of being a human – with every blink of my eyes I am “becoming less special” as computer programs with less and less inductive bias master more and more complex tasks. We also see how human learns in a way that’s not too different from machine sometimes (see also this paper).

I interpret this breakthrough as new evidence revealing the nature of our very miraculous existence, altough we might still be several lightyears far from reaching any conclusion. Still, I believe we are in this new age of enlightment, where we are making unprecedented progress in discovering our ordinariness (in contrast to the 18th century, when we were discovering our extrodinariness). The only thing we could do is making progress rather than finding the answer.