GPT-3 is (the process of) general AI being born in front of our eyes, catching us unprepared, making us realize that the we may be a bit rigid in understanding of what intelligence really means. If primed correctly, can GPT-3 produce worthwhile improvements to its own paper or implementation? Can GPT-3 be made to prime and query itself in a loop, by asking it to produce a better query for itself? Can its knowledge be already used to disrupt existing frontiers of human thought? The simple and safe answer here is that we don't know yet.

In a world full of ideological adversity and cold-blooded corporate competition, slowing down the release of technology which may or may not be capable of a runaway effect (a.k.a. technological singularity) in the hands of an adversary is an intelligent and prudent thing to do.