Jack Clark

To many, the ostensibly imminent automation of AI research looks sufficient for the rapid subsequent development of other world-changing capabilities.

It would constitute, in Jack Clark's words, "the crossing of a Rubicon into a nearly impossible to forecast future." The writers of AI 2027 think the full automation of coding will produce systems vastly better than the best humans at every cognitive task within just one year.

On the extreme version of this view, as Ajeya Cotra puts it, it'll take only months or weeks or days to create sci-fi technologies that would have taken centuries at human rates.

Perhaps most importantly, this progress could unfold almost entirely within the confines of data centers, unconstrained by the rate of AI's wider diffusion into the economy and take the rest of the world by surprise.