> And yet, if I had only said, “there is no way that online video will meaningfully contribute to economic growth,” I would have been right.
online video conferencing saved parts of the economy from shutdown during COVID. The trend seems somewhat persistent post-pandemic and may save lots of time commuting/traveling
online education has grown enormously but I‘m not sure this would get picked up in GDP numbers. I’m India right now and almost everyone on a commuter train watched a video lecture on their phone. I’ve fixed multiple things in my home thanks to ‚youtube university ’ for which i would have needed to hire a professional otherwise.
I wonder about the implicit assumption that AI is neutral wrt advancing progressive policy.
For example, you could probably use LLMs today to mass produce letters of support for specific developments. Very soon, they'll be eloquent and persuasive as well.
You'll be able to do the same to argue for better legislation, and eventually automate CBO-style scores that are both more specific and more robust.
Of course, this could become an arms race, but I would hope that the superior facts of the progressive policies would mean it is an uneven one.
Historically it has looked quite intractable, but I'd say that's changing recently and may spur more grants.
The bigger problem though is it's a first-world issue, so you automatically get a big haircut on cost effectiveness. Even so, this is one of only a handful of things that are prioritized for those geos.
I think looking at policy specifically can be first world and narrow, but I’m with a company looking to raise the accuracy of property assessment period, and I think there’s a lot of impact to be had there.
For instance we’ve worked with some poor Eastern European countries who are only now getting proper data collection and models set up, which will help them tax much more effectively, as well as help stabilize their real estate market. It’s still early days though!
From the post we don't get information about the acceleration rate of AI capabilities but on the impact on the economy. This argument is thus against slow takeoff with economic consequences but not against slow takeoff without much economic consequences.
So updating from that towards a discontinuous takeoff doesn't seem right. You should be updating from slow takeoff with economic consequences to slow takeoff without economic consequences.
There will be a complete 4 year interval in which world output doubles, before the first 1 year interval in which world output doubles. (Similarly, we’ll see an 8 year doubling before a 2 year doubling, etc.)
At some point there will be incredibly powerful AI systems. They will have many consequences, but one simple consequence is that world output will grow much more quickly. I think this is a good barometer for other transformative effects, including large military advantages.
I believe that before we have incredibly powerful AI, we will have AI which is merely very powerful. This won’t be enough to create 100% GDP growth, but it will be enough to lead to (say) 50% GDP growth. I think the likely gap between these events is years rather than months or decades.
In particular, this means that incredibly powerful AI will emerge in a world where crazy stuff is already happening (and probably everyone is already freaking out). If true, I think it’s an important fact about the strategic situation.
Though there are other takeoff-ish questions that are worth discussing, yeah.
Thanks for this clarification! I guess the "capability increase over time around and after reaching human level" is more important than the "GDP increase over time" to look at how hard alignment is. It's likely why I assumed takeoff meant the former. Now I wonder if there is a term for "capability increase over time around and after reaching human level"...
Yes. In a slow takeoff scenario where we have AI that can double GDP in 4 years, I don't regulations will stand in the way. Some countries will adopt the new technologies and then other countries will follow when they realize they are falling behind. NIMBYs and excessive regulations are a problem for economic growth when GDP is growing by 2% or 3% but probably won't matter much if GDP is growing by 20% or 30%.
I like the way you summarized this, thanks!
You're welcome.
Happy to be useful!
I wonder about the implicit assumption that AI is neutral wrt advancing progressive policy.
For example, you could probably use LLMs today to mass produce letters of support for specific developments. Very soon, they'll be eloquent and persuasive as well.
You'll be able to do the same to argue for better legislation, and eventually automate CBO-style scores that are both more specific and more robust.
Of course, this could become an arms race, but I would hope that the superior facts of the progressive policies would mean it is an uneven one.
I like that housing is one of the big things mentioned that AI will not change. I’m consistently surprised by how low EA rates land-use reform.
I may just not be well-run enough, but are there any major EA organizations or donations going towards land policy?
Open Phil grants quite a bit to land use reform https://www.openphilanthropy.org/grants/?q=&focus-area[]=land-use-reform
Thanks for the link!
Edit: looks like the total is around $20 mil to land use reform. Still a good amount but I’d think it would be a higher priority.
Historically it has looked quite intractable, but I'd say that's changing recently and may spur more grants.
The bigger problem though is it's a first-world issue, so you automatically get a big haircut on cost effectiveness. Even so, this is one of only a handful of things that are prioritized for those geos.
I think looking at policy specifically can be first world and narrow, but I’m with a company looking to raise the accuracy of property assessment period, and I think there’s a lot of impact to be had there.
For instance we’ve worked with some poor Eastern European countries who are only now getting proper data collection and models set up, which will help them tax much more effectively, as well as help stabilize their real estate market. It’s still early days though!
I’d like to think proper land management is also crucial going from third or second world to first world, as Scott points out in his review of How Asia Works. https://astralcodexten.substack.com/p/book-review-how-asia-works
From the post we don't get information about the acceleration rate of AI capabilities but on the impact on the economy. This argument is thus against slow takeoff with economic consequences but not against slow takeoff without much economic consequences.
So updating from that towards a discontinuous takeoff doesn't seem right. You should be updating from slow takeoff with economic consequences to slow takeoff without economic consequences.
Does that make sense?
Paul Christiano operationalizes slow/soft takeoff as:
Though there are other takeoff-ish questions that are worth discussing, yeah.
Thanks for this clarification! I guess the "capability increase over time around and after reaching human level" is more important than the "GDP increase over time" to look at how hard alignment is. It's likely why I assumed takeoff meant the former. Now I wonder if there is a term for "capability increase over time around and after reaching human level"...
I guess I don't understand how slow takeoff can happen without economic consequences.
Like takeoff (in capabilities progress) may still be slow, but the impact of AI is more likely to be discontinuous in that case.
I was probably insufficiently clear on that point.
Yes. In a slow takeoff scenario where we have AI that can double GDP in 4 years, I don't regulations will stand in the way. Some countries will adopt the new technologies and then other countries will follow when they realize they are falling behind. NIMBYs and excessive regulations are a problem for economic growth when GDP is growing by 2% or 3% but probably won't matter much if GDP is growing by 20% or 30%.
I love this post, and I'm not convinced by some of the counters in the comments. (e.g., I don't think LLMs will help persuade anyone of anything.)
This Ezra Klein podcast is really good, if you haven't heard it.