Some people think GDP is a good metric for AI timelines and takeoff speeds, and that the world economy will double in 4 years before the start of the first 1-year doubling period, and that AGI will happen after the economy is already growing much faster than it is today.
Other than AGI, what technologies could significantly accelerate world GDP growth? (Say, to a doubling period of <8 years, meaning the whole world economy grows 9%+ per year, significantly faster than the fastest-growing countries today.)
I find myself struggling to think of plausible answers to this question. Here are some ideas:
--Cheap energy, e.g. from solar panels or fusion
--Cheap resources, e.g. from asteroid mining, undersea mining, automated mines...
--Robots and self-driving cars make transportation and manufacturing cheaper
--3D printing? Idk.
--Narrow AI? Seems like the most plausible answer, but narrow AI doing what, exactly? Driving cars? Manufacturing things? Already discussed that. Inventing new products? OK, but in that case won’t they also invent AGI?
My problem is that while all of these things seem like they could be a big deal by ordinary standards, they don’t seem like that big a deal. Looking back over US economic history, it seems to my quick glance that growth rates haven’t changed much in 200 years. (!!!) But over that time energy, resources, etc. have gotten lots cheaper in the USA, and all sorts of new tech has been developed. Worldwide, it looks like the last time annual GWP growth was less than half of what it is now (excluding the Great Depression) was… 1875! (At least according to my data, would love to see a more thorough investigation of this). The world looked hella different in 1875 than it does now in 2020; doubling world GDP growth rates again seems like a pretty tall order. I believe that AGI could do it, but what else could?
I think past acceleration is mostly about a large number of improvements that build on one another rather than a small number of big wins (as Katja points out), and future acceleration will probably be more of the same. It seems like almost all of the tasks that humans currently do could plausibly be automated without “AGI” (though it depends on how exactly you define AGI), and if you improve human productivity a bunch in enough industries then you are likely to have faster growth.
I expect “21st century acceleration is about computers taking over cognitive work from humans” will be the analog of “The industrial revolution is about engines taking over mechanical work from humans / beasts of burden.”
From that perspective, asking “What technology short of AGI would take over cognitive work from humans, and how?” is analogous to asking “What technology short of a universal actuator would take over mechanical work from humans, and how?” The answer is just: a bunch of stuff that’s specific to the details of each type of work.
Thoughts on some particular technologies, kind of at random:
I think that most of that automation is likely to involve new software, and so the size of the software industry is likely to grow a bunch. Increasing productivity in the software industry (likely via ML) would then be an important driver of productivity growth despite software currently being a small share of GDP.
I think that cheap solar power, automation of manufacturing and construction (including manufacturing industrial tools and construction of factories), and automation of service jobs are also very important stories.
I think that west probably could be growing considerably faster even without qualitative technological change, so part of the story may be western countries either getting out of their current slump or being overtaken.
The other part of your post is about how much qualitative change would correspond to a doubling of growth rates. I think you are moderately underestimating the extent of historical acceleration and so overestimating how much qualitative change would be needed:
I think the US over the last 200 years is a particularly bad comparison because at the beginning of the period it was benefiting a lot from colonization. Below I talk about the UK which I think is probably more representative. I chose the UK as the the most natural frontier economy after the industrial revolution, but I expect the exercise would be similar for other countries without complications.
Looking at growth over the last 200 years hides the fact that there was a period of more rapid acceleration followed by a stagnation. If we instead compared 1800 to 1950 we’d see a larger change in growth rates accompanied by a smaller qualitative change. So that’s probably more useful if you are looking for an existence proof (and I think low levels of current growth likely make acceleration easier).
In 1800 the US was growing rapidly in significant part because colonists were still taking new land and then increasing utilization of that land. So over the last 200 years you have a decrease in some kinds of growth and an increase in others. I don’t know much about this and it may be completely wrong, but given that the US was growing so much faster than the rest of the world and that there’s such a simple explanation that seems to check out, that’s what I’d assume is going on. If that’s right then it can still be OK to use the US as an example but you can’t use raw growth numbers to infer something about technological change.
If you want to see what’s happening in frontier economies since the industrial revolution then it seems more natural to use something like per capita GDP in the UK. If I look up the GDP per capita in the UK time series at Our World in Data and turn that into a graph of (GDP per capita growth rate) vs (time), I get:
So it seems to me like things really did change a lot as technology improved, growing from 0.4% in 1800-1850, to 1% in 1850-1900, to .8% in 1900-1950, to 2.4% in 1950-2000. What we’re talking about is a further change similar in scope to the change from 1800 to 1850 or from 1900 to 1950.
(I don’t know if there are other reasons the UK isn’t representative. I think the most obvious candidate would be that 1900-1950 was a really rough period for the UK, and then 1950-2000 potentially involves some catch-up growth.)
Thanks! This is my new favorite answer. I consider it to be a variant on Abram’s answer.
--I think the large number of small improvements vs. small number of large improvements thing is a red herring, a linguistic trick as you say. There’s something useful about the distinction for sure, but I don’t think we have any major disagreements here.
-- Re: “21st century acceleration is about computers taking over cognitive work from humans” will be the analog of “The industrial revolution is about engines taking over mechanical work from humans / beasts of burden.” Yes, this sounds like a good hypothesis to me. My objection to it right now is that software has already been a thing for seventy years or so. We’ve been automating lots of human cognitive work every decade for almost a century. Yet growth rates have gone down rather than up. So it seems to me that if software is going to make growth rates go up a lot, we have to say something extra to explain why it hasn’t done so yet. To be clear, I don’t think there is nothing to say on this; I would actually bet that someone has a decent explanation to give. I’m interested to hear it. (One thought might be: Steam engines started picking up steam with Watt’s engine in 1775, but growth in the UK stayed on-trend until about a century later! So maybe these things just take time.)
--And this objection is more general than that; it’s the stick I’m using to beat up everything right now. Yeah, I can imagine lots of little improvements causing an increase in the growth rate. It’s totally happened before. But for the last fifty years at least, it hasn’t been happening: There have been lots of little improvements in every sectory and some big improvements even, and loads of jobs have been automated away, but overall growth rates haven’t gone up, and have even declined. So, what reason do we have to expect that in the future, as even more improvements are found and even more jobs are automated, growth rates will go up? (Again, I would bet that there are good reasons, I just don’t know what they are right now)
--Re: From that perspective, asking “What technology short of AGI would take over cognitive work from humans, and how?” is analogous to asking “What technology short of a universal actuator would take over mechanical work from humans, and how?” I love this comparison. I totally agree that in principle we could automate everything or almost everything with narrow AI (incl. software), rather than AGI. After all, instead of building humanoid robots we built more specialized machines. However, I think that AGI is likely to come before we’ve automated everything with narrow AI, and moreover I think that if we were to automate everything with narrow AI, AGI would come very shortly thereafter, and indeed probably before we finished automating everything with narrow AI. (Because building AGI doesn’t require sampling from all jobs in the economy, but only a subset; we could automate those ones and then get AGI before we’ve automated the bulk of the jobs in the economy.) By contrast, the parallel argument about universal actuators isn’t as plausible. Having lots of really specialized actuators doesn’t help you much in getting a humanoid robot, because the bottleneck is finding the right design rather than getting stuff from one position/location to another. (Whereas for AGI, the bottleneck is finding the right design, and that’s the sort of cognitive task we are automating with narrow AI)
--Further thoughts on the analogy: I am fairly convinced that there are many important tasks which can be done most competitively by agent-like systems who are fairly general intelligences. (Perhaps the analogous thing for actuators is: There are some tasks that are better done by “General vehicles” which aren’t tied to a particular location, can travel over a large range of terrain types, and can transport a wide range of cargoes and also perform other tasks like digging and pulling things. i.e. pickup trucks. (extending the analogy further, perhaps pre-trained unsupervised world-models are like the engines that get mass-produced and put in cars, trucks, airplanes, tanks, and sometimes even fixed locations. So maybe engines are like universal actuators after all, in a sense.) And the classic problems of AI risk arise from these sorts of systems. So the question is how much relative progress will be made at automating these tasks vs. automating all the other tasks. If it’s very little, such that these are the last tasks to be automated to a significant extent (if CEOs and generals and researchers etc. are the last jobs to go!) then yeah the economy might be growing fast by the time classic AI risk concerns start to materialize. If however these jobs are automated at a similar or greater rate, then AI risk concerns will be materializing at the same time as the tech to accelerate the economy is invented, which means slightly before the economy actually accelerates.
Fair enough, thanks for the input and the data. In particular:
This neatly gets down to business. The issue becomes: We’ve seen doublings of growth rate (and halvings) a time or two in the past two centuries, so it’s reasonable to expect more in the next. And insofar as the explanation for these changes in the past was “Lots of things got better across all sectors of the economy” then we should take seriously the corresponding prediction for the future. But insofar as the explanation for these changes was e.g. “Yeah lots of things got better, but to a first approximation the main drivers of progress were Engines + electricity + …” then we should expect any future changes to come along with a similar list of the main drivers of progress. And then the question is: What would those drivers be? And the answer would probably be: Software/NarrowAI. And then the question would be: OK, but we’ve had software/NarrowAI for a while, why hasn’t it had an effect yet? And the answer would be… well, I don’t know what it is yet but I’m reasonably confident there is one.
--I agree GDP per capita in frontier economies is a more relevant metric than GWP. Why don’t we use that instead of GWP?
France is the other country for which Our World in Data has figures going back to 1400 (I think from Maddison), here’s the same graph for France:
There is more crazy stuff going on, but broadly the picture looks the same and there is quite a lot of acceleration between 1800 and the 1950s. The growth numbers are 0.7% for 1800-1850, 1.2% for 1850-1900, 1.2% for 1900-1950, 2.8% for 1950-2000.
And for the even messier case of China:
Growth averages 0 from 1800 to 1950, and then 3.8% from 1950-2000 and 6.9% from 2000-2016.
I don’t know what post to link to, but I recall at some point Robin Hanson articulated fully automated manufacturing as his guess about the next big bump in GDP doubling times.
The argument as I recall it:
Full automation means automation of everything including building the factories themselves.
Full automation plausibly requires advanced AI, but not full human-level AGI. So (especially if we believe in relatively slow AI progress) we might expect to see this significantly before an AGI-based boom.
Fully automated manufacturing would make manufacturing much cheaper by cutting out the human cost, and automated manufacture of factories would allow rapid scaling, and rapid responses to economic demands, which would be dramatic and game-changing. Production cycles (from idea to prototype to hitting the market) would be dramatically shortened.
Thanks, this is my favorite answer so far. [EDIT: Now Paul’s is my favorite.] It’s sorta what I had in mind with my list of candidates above. I guess your points #2 and #3 are the ones I’m skeptical of.
Re point 2: If we’ve automated everything, including building factories and entire production cycles, (a) doesn’t that involve figuring out how to make computers do a huge variety of tasks, many of which are quite intellectually difficult, such that plausibly the easiest way to do this is to create general AI rather than loads of narrow AIs (or else R&D tools that automate the automation process for us, but in that case they’d probably also help us get AGI) and (b) even setting aside that, couldn’t we string all these narrow AIs together to make an AGI?
Re point 3: Haven’t we come close to automating everything multiple times in the past century? I don’t know, but I would guess that 90%+ of the industrial/manufacturing jobs done by humans even just 50 years ago are now done by machines. But this 90% automation didn’t lead to a near-doubling of GWP growth rates.
90% automation only gives a ~10x increase in per-worker productivity in manufacturing. Since manufacturing is only a fraction of GWP, a 10x productivity increase only makes GWP (per capita) a few times larger. Take humans out of the process completely and the bottleneck is gone. The feedback loop is only constrained the availability of resources.
Good point. I guess I just find it implausible that humans will be COMPLETELY out of the loop prior to AGI. Some parts of the loops involve agenty tasks in which you draw on general world-knowledge to make novel plans and strategies and then execute them learning and adapting constantly.
I’m not sure if the agenty tasks you have in mind are considered part of manufacturing per se or business management. My impression from above is that production work and factory construction is being automated but design/engineering and business management are not. I’m not sure, but it does seem likely that humans could be out of the loop without AGI. (Though of course AGI could happen before narrow AI actually achieves this level in practice).
...there are somewhere between six and ten billion people. At any given time, most of them are making mud bricks or field-stripping their AK-47s. - Neal Stephenson, Snow Crash
When we think of new technologies, we typically think of expensive, high-tech innovations, like energy production, robotics, etc. I would suggest that broader adoption of existing technologies, including social technologies, would have a bigger global impact.
For example, one technology that could dramatically impact GDP is improved managerial technology. This paper describes a study of this in India. Among the findings in the paper (or in references that it cites):
100% productivity spreads between the 10th and 90th percentile in US commodity-producing firms
A ratio of the 90th to the 10th percentiles of total factor productivity is 5.0 in Indian and 4.9 in Chinese firms
After improving management in the studied firms, “We estimate that within the first year productivity increased by 17%; based on these changes we impute that annual profitability increased by over $300,000. These better-managed firms also appeared to grow faster, with suggestive evidence that better management allowed them to delegate more and open more production plants in the three years following the start of the experiment”
FWIW, world GDP growth rates have if anything been decreasing over the last ~80 years
Interesting. Yeah, I guess if the less-developed world suddenly adopted cutting-edge tech and practices, that would be enough of a boost to grow at 9%+ for a few years until they caught up to the developed countries and slowed down to developed-country rates.
What could cause that to happen, though? Shouldn’t we expect the diffusion of cutting-edge tech and practices to take place over several years (decades, even) in the absence of AGI?
China had 40 years of that kind of growth. For poor African countries there’s more then a few years of growing like that to catch up.
If someone manages to solve online teaching in a scaleable way it might be able to change the way people in developing countries run their businesses in shorter amounts of time.
Yeah, someone else suggested a novel nootropic drug as one answer—online education is basically an alternative form of that drug that is easier to realize (or at least, it’s hard is a very different way)
Excluding AI, and things like human intelligence enhancement, mind uploading ect.
I think that the biggest increases in the economy would be from more automated manufacturing. The extreme case is fully programmable molecular nanotech. The sort that can easily self replicate and where making anything is as easy as saying where to put the atoms. This would potentially lead to a substantially faster economic growth rate than 9%.
There are various ways that the partially developed tech might be less powerful.
Maybe the nanotech uses a lot of energy, or some rare elements, making it much more expensive.
Maybe it can only use really pure feedstock, not environmental raw materials.
Maybe it is just really hard to program, no one has built the equivalent of a compiler yet, we are writing instructions in assembly, and even making a hello world is challenging.
Maybe we have macroscopic clanking replicators.
Maybe we have a collection of autonomous factories that can make most, but not all, of their own parts.
Maybe the nanotech is slowed down by some non-technological constraint, like bureaucracy, proprietary standards and patent disputes.
Mix and match various social and technological limitations to tune the effect on GDP
I don’t think any one of those would have an impact anywhere near that big. If nothing else, they’ll each take long enough to mature that their impact will be spread out over many years. However, I would suggest the combination of distributed, autonomous manufacturing (of which 3D printing is one part), generative design, and materials informatics (and related informatics technologies) could get there.
Right now bringing new stuff to market and scaling it up is many times faster and cheaper for software than anything else. The more you can reduce the resource burden, and with it the number of people and organizations that have to buy in, to turning ideas into products, the smaller that gap can get. I would naively assume that this is analogous to any other catalyst—lower the energy barriers for the rate-limiting steps, and you get exponential speed-up. It also reduces the cost of entry and cost of failure, making it possible for many more people to participate in innovation.
One problem with this idea is that what I’m proposing is to essentially commoditize scale-up, manufacturing, and some parts of R&D, to make them nearly free. I’m making no attempt to work out the downstream effects of that on the service sector, which is a majority of GDP. I do think that’s likely more a measurement problem with GDP (if a problem at all), though, and that the kind of improvement I’m suggesting could easily lead to a less-easily-measured acceleration of economic value growth.
Digital matter, in other words, molecular manufacturing.
With a Star Trekian autogenous home synthesizer, one could expect Moore’s Law-like growth.
I agree this would do it. However it seems unlikely to come prior to AGI, and also if it does come before AGI I think someone would use it to create AGI before the first 9%+ GWP growth year. (How? By building loads of awesome cheap computers, and then brute-forcing AGI with the extra orders of magnitude of compute. If a sugar cube can do 10^21 flops, then a datacenter can do 10^26 I guess (unsure), and that means 10^32 operations in ten days or so.)
Do you expect pre-takeoff AI to provide this? What sort of AI and production capabilities are you envisioning?
Or are you answering this question without reference to AI? If so, what would make this useful for estimating AI timelines?
There is nothing magical about it, so yes—AI will help, but humans are enough. I would expect a sustained, 100s of millions/$1bn 10 year effort would bring us a lot closer to this hardware technological ‘maturity’ (we are not even trying hard now—a mole of carbon-12 has a mass of 12g and it contains 6*10^23 atoms, yet there is nothing stopping us from building small machines with mere thousands/millions of atoms). Obviously you could say that this sort of money would help with anything, but I believe it would be one of the best value/$ projects).
With regards to molecular manufacturing, one could imagine a multitude of ideas that are perfectly feasible from classical physics standpoint yet nowadays are in the realm of sci-fi—examples including humanoid robots built bottom-up (with milliond of tiny motors, moving more majestically than a human), mechanical computers the size of a sugar cube with 10^21 FLOPS and approaching Landauer’s principle (~computronium—https://youtu.be/yVX9Ob4SjGA) or as I mentioned previously, an autogenous replicator which would allow any object to be replicated, including itself (you know what the curve looks like...)
The second example ties with the AI/AGI—you do not have to worry about Moore’s law in the narrow sense, even in terms of FLOPS/$ as current semiconductor substrate in simply a local maximum. Regardless of whether you are in the ‘scaling’ camp or not, more FLOPS would surely help test this hypothesis as well as many others… No one knows the hurdles in front of us, but it would surely ‘help’ to build AGI—obviously in quotation marks given the risks.
A tricky thing here is that it really depends how quickly a technology is adopted, improved, integrated, and so on.
For example, it seems like computers and the internet caused a bit of a surge in American productivity growth in the 90s. The surge wasn’t anything radical, though, for at least a few reasons:
Continued technological progress is necessary just to sustain steady productivity growth.
It’s apparently very hard, in general, to increase aggregate productivity.
The adoption, improvement, integration, etc., of information technology was a relatively gradual process.
If we instead suddenly jumped from a world where no company has information technology to one where every company is using 2020-level information technology (and using it with 2020-level tacit knowledge, IT-enabling capital investments, IT-adapted business practices, complementary technologies, etc.), then the productivity growth rate for that year would probably have been very high. But the gradualness of everything flattened out the surge. Given how slowly diffusion happens globally, I actually wouldn’t be surprised if the surge was totally invisible at the global level.
So if we want to predict that some technology (e.g. fusion power) will help surge the growth rate above some high threshold, we will also typically need to predict that its aggregrate impact will be unusually sudden.
How about a novel nootropic drug? Or advancements in neurolink technology to enhance brain-computer interfaces? Plus other biological and medical advancements that one could conceive of. Weight bearing suits that ease repetitive stress and increase lifting capacity. Personal climate controlled suits that make you comfortable anywhere.
I’ve made the argument in a post I wrote that we will see the first genetically engineering humans in the next decade. The technique will likely be embryo selection, which results in somewhat modest trait gains even with hundreds of embryos. However, the gains made by massive embryo selection (hundreds to thousands of embryos rather than ten) are still likely to exceed the effect of even the strongest nootropics.
I do find it plausible that if we had some sort of cheap drug that made the user +2 SD’s in IQ while also being more focused and creative, that would kick the economy into a much higher growth rate. Seems unlikely to happen though. Maybe narrow AI for bio R&D could do it. Maybe I’m overestimating how much it would take—maybe only half an SD would do it?
Cloning. Increase world population by 9% per year and you’ll add 9% in GDP fairly easily.
Which also points to GDP per capita being a more interesting measure?
OK, thanks, I hadn’t thought of that! However, this seems very unlikely to happen anytime soon. In fact it basically can’t happen for the next 25 years at least.
Probably true. And even then I don’t see any reason why there would be an interest in increasing the world population by that much. But it’s a technological possibility...
Following this logic, couldn’t a government just drive GDP growth by paying people to have more children?
Right now, it seems that one can get a $1000 tax credit per child under the age of 17 in the US. The GDP per capita in the US is about $65,000. Therefore, it seems that there is lots of room to increase GDP by paying even more per child.
Having more children the natural way would do the trick, but that’s not a technological solution, which was what was being asked for?
But again, this “only” increases GDP, not GDP per capita. It wouldn’t actually change anybody’s living standards.
China had a longer time that kind of growth through having good governance and effective economic policy.
If a technology such as the seeing rooms that Cummings talked about suddenly allow decision making and governance to improve by a large margin, we might get that kind of economic growth.
The fact that the cost of building subway tunnels increased rather then decreased isn’t because energy prices or resource prices rose but because of a mix of lack of innovation and complex regulation.
How about 3D bioprinting as a form of regenerative medicine?
It’s been said that about half of all people have an IQ less than 100. Some psychologists have pointed out that those with IQs less than 90 have a difficult time finding good work in advanced knowledge-driven economies, and manual labor has been either exported to other countries or replaced with robots, leaving part of the labor pool underutilized.
So the shape of the idea that would generate 9%+ GDP growth is a set of technologies and/or political configurations that bring people of all IQs enthusiastically into the labor force. Not just employment opportunity, but situations that would be gleefully embraced, and productive, regardless of IQ. Work that is useful and fulfilling and worth doing for all involved.
This is not the actual idea, of course, only a statement of what might be its shape. If I actually had that idea, I would be talking to venture capitalists at this moment rather than typing this comment.
This would be good only for a temporary gain of three to six years of high growth. After that, we would be at full employment, and although the indirect gains would likely flow for a long time, national growth would likely fall below the 9% figure.
What’s more important, though, is that such an outcome (bringing gleeful and productive employment to many) would make a lot of people happier, regardless of GDP growth.