Artificial Intelligence's Rise Means Change Will Happen Even Faster

tech.hindustantimes.com:

 So what did we learn this year? Specifically, I am asking how the real-world events of the past 12 months should have caused us to update our social scientific views. In a year rich with news, which insights turned out to be more useful, or less? The first big lesson from 2023 is hopeful: Huge gains in technology can be made very quickly.

A year and a few months ago it was not obvious to most people that artificial intelligence, in particular GPT-4, would be so transformative across so many dimensions. Now that view has become conventional wisdom. In fact, expert opinion now expects AI to surpass humans in most fields of intellectual endeavor in less than 10 years. I know many experts in the field who think it will be in less than five years. By some metrics, and at great cost, AGI (artificial general intelligence) might even be possible right now.

So we have learned that what may prove to be one of the most important advances in human history can sneak up on most people in little more than a year. We thus need to update our thinking about whether major innovations can come seemingly “out of nowhere,” following periods of relative stagnation. (In similar fashion, in early 2020 many experts thought a Covid vaccine would take years to develop, when in fact mRNA vaccines were effective and available in less than a year.)

Theories of sudden leaps should therefore rise in status. Yes, these breakthroughs depend in part on the slow accretion of expertise and execution — but then they burst onto the scene.

Another takeaway from 2023 is more depressing: Deterrence is less powerful than I thought. Persistence, combined with a belief in one's cause, is worth more.

The Israeli military is much stronger than Hamas, for example, and is