Nuclear knowledge diffused because its discovery hit a narrow window

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

1. 4:24

   um i want to start with neil um because your your most recent novel polistan uh harkens back to i think probably the last period in history where there is a comparably uh significant feeling of acceleration like similar g-forces which is the dawn of the nuclear age um and i i wonder were Were you thinking about the current moment, or did it affect the way that you approached that particular project?

2. 4:54

   And how do you think that those moments are similar or different? Yeah, it wasn't a conscious decision to draw a parallel between those two eras. But the parallel is there regardless, as you point out. And, you know, I'm pretty deep into that, into the timeline right now of how the understanding of nuclear fission got started and how it propagated from there. And it actually got started in a ski trip between Lisa Meitner and her nephew in Sweden when they were trying to make sense of some results out of Germany in 1938 and spread to a small number of people in the physics community before the war started and everything got shut down.

3. 5:51

   So if it had happened any earlier, nobody would have bothered making tanks and airplanes in World War II because everybody would have been making atomic bombs. If it had happened any later, it would have been confined to just a tiny group of people who would have been immediately shut down or sequestered by their government. But instead, it was in this kind of Goldilocks place where enough people knew about it from different countries that the knowledge of it kind of slowly diffused, and it became the dominant thing that people worried about during the war leading up to the Manhattan Project.

4. 6:35

   So, you know, how that applies to the AI situation right now, maybe it's something that we could discuss instead of me just lecturing about it. But, you know, I'll say that a lot of people were taken by surprise by the rapid onset of what AI could do and how quickly it diffused into the general awareness. And in some respects, I think we're in a little bit of a similar situation now in that some people are, in effect, asking, you know, why work on anything but AI?

5. 7:21

   Does AI make everything else kind of irrelevant and not worth bothering with? Or, you know, should we continue to pursue old school technology while we wait for the dust to settle? I mean, I want to follow up on that a little bit. I think it's interesting to compare the atomic age with this moment. But I think there's another way to look at it, which is a little bit more cautionary. I would say that for many of us, the atomic age has played out to be a disappointment.

6. 7:57

   What I mean by that is this. If you read the sci-fi written at the time, at the dawn of the atomic age, it was seen as a transformative technology. Like it would transform every aspect of society. The way it played out is quite different. I mean, on the strategic level, it became a tool of stasis. It essentially maintained spheres of influence and stopped wars in a way that previous technologies were only envisioned to do but did not.

7. 8:27

   That was not quite how people envisioned it would work out. And on the civilian side, I think all of us as young children, when we learned how nuclear power plants actually work, we were disappointed, I think. At least I was. Here you have this amazing, powerful physics thing. And the way you generate power is you turn water into steam and spin some turbines. I don't know. I was disappointed when I learned that that was how that worked.

8. 8:56

   I was like, this is really dumb. And that's basically where we are. So there's a huge amount of disappointment, I think, psychically over the atomic age, not living up to its promises. Now, is that going to happen to the AI age? I'm not entirely sure. I don't know how it's going to play out. To me, the most useful applications of AI so far of large language models has been in the realm of protein folding, right?

9. 9:27

   These are large language model-like techniques being applied to protein folding and essentially solving a problem that was deemed unsolvable. That, to me, is the most impactful thing AI has done so far. So if AI ends up being, you know, the sort of technology that leads to hundreds or thousands of similar level kinds of innovations, then I would say this is way, way more impactful than the atomic age.

10. 9:58

    But at the same time, another possibility is we just end up with AI replacing a bunch of semi-creative jobs and leading to a kind of industrial revolution 2.001. which is not all that promising. I would feel about that, about the same way I feel about nuclear power plants. So that's kind of how I feel. We don't know how it's going to turn out.