I have a lot of thoughts but not a lot of time, so apologies if this is a bit scatterbrained.
I’ve read your blog, Roodman’s blog from last year and a lot of Roodman’s report. I see this line of thinking in the following way:
Some EAs/rationalists/AI alignment groups believe that AI could be transformative because AI itself is unlike anything that has come before (I mostly share this view). Your and Roodman’s line of inquiry is to do a check on this view from an “outside view” perspective, using long term economic growth to make sure that there is more supporting the transformative AI idea than just “AI is totally different from anything before”—the inside view.
This could be particularly useful for AI timelines, and also perhaps convincing skeptics of transformative AI that this is worth considering.
The big problem then, is that economic growth over the last century+ has been at a fairly constant rate, or at least, it’s certainly not increasing.
So I completely agree with your assessment of Roodman’s model.
I actually wrote about this is a blog post earlier this year. I’m interested in the question of whether we should focus any time thinking about economic growth as a major policy outcome; if transformative AI is very close, then getting US GDP from 1.5% to 3% is kinda unimportant.
If I’m an AI skeptic, I don’t think Roodman’s model convinces me of much. It can’t really rely on GWP data of the last century because it doesn’t fit the model, so the entirety of the argument stems from GWP estimates going back 10,000 years. And the best guesses of late 20th century economists about how to measure “Gross production” in 5000 BC just seem really shaky.
So, yes, it seems really unconvincing to pull AI timelines from this data.
Also, on endogenous growth models, you don’t exactly mention it in your post, but what really jumped out at me was that you say around ~1880, people started getting richer, there wasn’t just population growth when technological progress was made. But then the next point seems very clear: there’s been tons of population growth since 1880 and yet growth rates are not 4x 1880 growth rates despite having 4x the population. The more people → more ideas thing may or may not be true, but it hasn’t translated to more growth.
So if AI is exciting because AIs could start expanding the number of “people” or agents coming up with ideas, why aren’t we seeing huge growth spurts now?
Ok final thing: I think the question of GDP measurement is a big deal here. GDP deflators determine what counts as “economic growth” compared to nominal price changes, but deflators don’t really know what to do with new products that didn’t exist. What was the “price” of an iPhone in 2000? Infinity? Could this help recover Roodman’s model? If ideas being produced end up as new products that never existed before, could that mean that GDP deflators should be “pricing” these replacements as massively cheaper, thus increasing the resulting “real” growth rate?
It certainly seems plausible to me, but I’m not sure what the practical relevance is. Would this convince people that transformative AI is a possibility? Would it give us better timelines? It seems like we’re just kinda inventing our own growth model again and then declaring that shows an “outside view” that transformative AI is a possibility. This seems unlikely to convince skeptics, but perhaps the critique of GDP calculation alone is worth broadly articulating before making any claims about AI.
I think the question of GDP measurement is a big deal here. GDP deflators determine what counts as “economic growth” compared to nominal price changes, but deflators don’t really know what to do with new products that didn’t exist. What was the “price” of an iPhone in 2000? Infinity? Could this help recover Roodman’s model? If ideas being produced end up as new products that never existed before, could that mean that GDP deflators should be “pricing” these replacements as massively cheaper, thus increasing the resulting “real” growth rate?
This is an interesting idea. It wasn’t a focus of my work, but my loose impression is that when economists have attempted to correct for these kinds of problems the resulting adjustment isn’t nearly large enough to make Roodman’s model consistent with the recent data. Firstly, measurements of growth in the 1700s and 1800s face the same problem, so it’s far from clear that the adjustment would raise recent growth relative to old growth (which is what Roodman’s model would need). Secondly, I think that when economists have tried to measure willingness to pay for ‘free’ goods like email and social media, the willingness is not high enough to make a huge difference to GDP growth.
Thank you for this comment! I’ll make reply to different points in different comments.
But then the next point seems very clear: there’s been tons of population growth since 1880 and yet growth rates are not 4x 1880 growth rates despite having 4x the population. The more people → more ideas thing may or may not be true, but it hasn’t translated to more growth.
So if AI is exciting because AIs could start expanding the number of “people” or agents coming up with ideas, why aren’t we seeing huge growth spurts now?
The most plausible models have diminishing returns to efforts to generate new ideas. In these models, you need an exponentially growing population to sustain exponential growth. So these models aren’t surprised that growth hasn’t increased since 1880.
At the same time, these same models imply that if increasing output causes the population to increase (more output → more people), then there can be super-exponential growth. This is because the population can grow super-exponentially with this feedback loop.
So my overall opinion is that it’s 100% consistent to think:
The increased population of the last 100 years didn’t lead to faster growth
If AGI means that more output → more people, growth will accelerate.
I have a lot of thoughts but not a lot of time, so apologies if this is a bit scatterbrained.
I’ve read your blog, Roodman’s blog from last year and a lot of Roodman’s report. I see this line of thinking in the following way:
Some EAs/rationalists/AI alignment groups believe that AI could be transformative because AI itself is unlike anything that has come before (I mostly share this view). Your and Roodman’s line of inquiry is to do a check on this view from an “outside view” perspective, using long term economic growth to make sure that there is more supporting the transformative AI idea than just “AI is totally different from anything before”—the inside view.
This could be particularly useful for AI timelines, and also perhaps convincing skeptics of transformative AI that this is worth considering.
The big problem then, is that economic growth over the last century+ has been at a fairly constant rate, or at least, it’s certainly not increasing.
So I completely agree with your assessment of Roodman’s model.
I actually wrote about this is a blog post earlier this year. I’m interested in the question of whether we should focus any time thinking about economic growth as a major policy outcome; if transformative AI is very close, then getting US GDP from 1.5% to 3% is kinda unimportant.
If I’m an AI skeptic, I don’t think Roodman’s model convinces me of much. It can’t really rely on GWP data of the last century because it doesn’t fit the model, so the entirety of the argument stems from GWP estimates going back 10,000 years. And the best guesses of late 20th century economists about how to measure “Gross production” in 5000 BC just seem really shaky.
So, yes, it seems really unconvincing to pull AI timelines from this data.
Also, on endogenous growth models, you don’t exactly mention it in your post, but what really jumped out at me was that you say around ~1880, people started getting richer, there wasn’t just population growth when technological progress was made. But then the next point seems very clear: there’s been tons of population growth since 1880 and yet growth rates are not 4x 1880 growth rates despite having 4x the population. The more people → more ideas thing may or may not be true, but it hasn’t translated to more growth.
So if AI is exciting because AIs could start expanding the number of “people” or agents coming up with ideas, why aren’t we seeing huge growth spurts now?
Ok final thing: I think the question of GDP measurement is a big deal here. GDP deflators determine what counts as “economic growth” compared to nominal price changes, but deflators don’t really know what to do with new products that didn’t exist. What was the “price” of an iPhone in 2000? Infinity? Could this help recover Roodman’s model? If ideas being produced end up as new products that never existed before, could that mean that GDP deflators should be “pricing” these replacements as massively cheaper, thus increasing the resulting “real” growth rate?
It certainly seems plausible to me, but I’m not sure what the practical relevance is. Would this convince people that transformative AI is a possibility? Would it give us better timelines? It seems like we’re just kinda inventing our own growth model again and then declaring that shows an “outside view” that transformative AI is a possibility. This seems unlikely to convince skeptics, but perhaps the critique of GDP calculation alone is worth broadly articulating before making any claims about AI.
This is an interesting idea. It wasn’t a focus of my work, but my loose impression is that when economists have attempted to correct for these kinds of problems the resulting adjustment isn’t nearly large enough to make Roodman’s model consistent with the recent data. Firstly, measurements of growth in the 1700s and 1800s face the same problem, so it’s far from clear that the adjustment would raise recent growth relative to old growth (which is what Roodman’s model would need). Secondly, I think that when economists have tried to measure willingness to pay for ‘free’ goods like email and social media, the willingness is not high enough to make a huge difference to GDP growth.
Thank you for this comment! I’ll make reply to different points in different comments.
The most plausible models have diminishing returns to efforts to generate new ideas. In these models, you need an exponentially growing population to sustain exponential growth. So these models aren’t surprised that growth hasn’t increased since 1880.
At the same time, these same models imply that if increasing output causes the population to increase (more output → more people), then there can be super-exponential growth. This is because the population can grow super-exponentially with this feedback loop.
So my overall opinion is that it’s 100% consistent to think:
The increased population of the last 100 years didn’t lead to faster growth
If AGI means that more output → more people, growth will accelerate.