9 Comments

"Frequency of phase transitions per year"? Isn't that just a bunch of nonsense?

Expand full comment

Growth in a finite world often consists of "S-shaped" curves. We can see the bottom ends of many of these curves in areas where there appears to be exponential growth. It seems likely that many of these curves will exhaust themselves within a century. Transistor sizes can't keep getting smaller and smaller at anything like the current rate. The next century will likely be the one where lots of big and important S-curves get exhausted. Of course, there may still be other, slower S curves going on. We won't master stellar farming in 100 years, for example. Progress won't come to an end. However, that still makes the coming century look especially interesting, which in turn raises the issue of: how come we are witnessing it? Simulism offers one possible answer. Chance is another explanation. Maybe there are other possibilities.

Expand full comment

The same considerations that make us think the 21st century CE will be the most important so far make us suppose that later centuries will be even more important.

As for being early: The End of History is for the universe to be packed with life. If in our time the universe is almost empty of life, we see that we are in the first of the two historical stages—the relatively empty one. If the full stage will last as long as or longer than the empty stage, we are certainly early, compared with the average “person.” (*Person* = *rational being*.) But even if the full stage is relatively brief, it contains so many “people” that you and I will probably be early compared to the average “person.” Also, our time may fall in the earlier part of the empty stage, which would make us early even if the full stage were only momentary.

Expand full comment

But maybe there are big differences all the way up a tall ladder of options.

Expand full comment

Hmm, so I agree that we tend to *define* intellegent life as "life that is as intellegent as humans", so of course that would introduce an inherent bias into our choice of reference class.

But I also genuinely believe that there is big qualitative distinction between humans and whatever the next most intellegent species that has ever existed on Earth.

Expand full comment

Yes, I was primarily criticizing the "last seen tech is the most important ever" version of the argument. I'd also bet against seeing 50 doublings in the next century, though I feel much less confident there.

Expand full comment

You could separate out "most important century" from "this particular set of foreseeable choices and challenges faced by humans now is the most important ever."

I share your skepticism of the latter---there will likely be many important choices in our future.

But with respect to the century, it seems pretty likely to be the most important century, because we are likely to see enough acceleration that we approach technological maturity, and see a large fraction of all the growth that human civilization could ever see. We could very easily have ~50 doublings this century as we fill up the solar system, while humanity has a whole has had <25 so far and <100 to fill up the whole universe. So it seems like about a third of all growth is likely to happen over the next century, and if that happens I think it is quite likely to be the most important century. I would guess that other measures like political change will also be radically accelerated, though I don't even think that's necessary.

I would have expected you to roughly agree about likely future growth dynamics, I'm not sure if you meant to criticize the narrower view, if I'm wrong about your view, or if you mean this as an outside-view argument against the "sequence of accelerating growth modes" perspective. I don't find the outside view update very persuasive, since I think that everyone *can* just make the same statement. People in the industrial revolution could say this was the most important millennium, and our children can say this is the most important decade, and our crazy robot descendants can say that this is the most important month.

Expand full comment

According to many singularity models, we're actually barely on-time. Not early at all. Even when considering counterfactual earliness (earliness relative to ourselves, rather than to aliens), historical population data can be used as an analog to suggests that the singularity should have happened almost 2000 years ago.

https://uploads.disquscdn.c...

Expand full comment

great post. Love that it starts so small, and drives the logic out to infinity and beyond.

Expand full comment