Until 1700 the world economy did not really grow—it just stagnated. Over the previous 17 centuries global output had expanded by 0.1% a year on average, a rate at which it takes nearly a millennium for production to double. Then spinning jennies started whirring and steam engines began to puff. Global growth quintupled to 0.5% a year between 1700 and 1820. By the end of the 19th century it had reached 1.9%. In the 20th century it averaged 2.8%, a rate at which production doubles every 25 years. Growth has not just become the norm; it has accelerated.
Assume those loops have maximum force and the economy becomes “information produced by information capital, which is produced by information, which in turn is producing information ever faster every year”, as William Nordhaus, a Nobel laureate in economics, wrote in a paper in 2021. This brings about the “singularity”—a point when output becomes infinite. The singularity is really a counterargument: proof that the model must, eventually, be proved wrong. But even the first step on the journey, a big acceleration in growth, would be a profound event.
I think a more important question is not whether AI will make economic growth explode but rather who that economic growth will benefit and in general how those benefits will be distributed.
I'd rather just have a world where people move a little slower, care less about efficiency, appreciate the smaller things in life, and stop forcing endless upgrades of every kind on everyone with new phones, new apps, soulless art, and new ways of doing things. But that's just me.
You can have really massive growth in markets along with massively wide inequality at the same time among the human population. You can also now (for the first time in global history) have laws of supply and demand start to service non-human intelligence.
Future transactions can have bots on both ends, running an economy that continually grows to service the needs of (probably more efficient) machine capitalists rather than human capitalists.
Just like when human agriculture and industrialization broke multiple organically forming food webs in the natural world, so could it be with human industry
Some weakly held opinions... Cheaper goods and services, more wealth inequality, more power concentration in the hands of a few individuals (and a few nation states), more expensive scarce assets like land, more jobs where humanness is valued, more climate change (in the short-run), better medicine, warfare outcomes even less coupled with population counts, less human casualties in war although bimodally distributed, new arms race between major powers, unpredictable social and civic consequences.
When I ask people what they would like to change about their life, pretty much everyone mentions that their apartment is too small and too expensive.
Because everyone wants to live right in the center of a big city.
While at the same time, when I look at Google maps, big cities are tiny. I zoom out a bit and every big city is just a tiny speck in the middle of nowhere.
When construction becomes cheaper, it will be more compelling to build houses outside the city. And when driving in comfortable autonomous electric buses is available frequently and cheaply, living outside the city becomes more compelling.
I think two possible effects of AI are often conflated.
On the one hand you can imagine that work gets supercharged, allowing companies to produce 10x the number of widgets at 1/10th the cost. The economy would grow rapidly, wealth inequality would presumably be exacerbated, jobs would be automated, we might need some version of universal basic income, and so on. People debate whether or not this kind of transition is imminent or if it'd take decades.
On the other hand, it's conceivable that not much would happen in the "bulk" of the economy while at the same time the frontier of humanity might be pushed forward. We may see new treatments for diseases, new types of energy production, and so on. In this version of the world, jobs would mostly remain unchanged (at least in the short to intermediate term), perhaps with some small multiplicative efficiency factor, the economy wouldn't grow rapidly, there wouldn't be any mass unemployment, and so on.
In my mind, I'm much more excited about the second kind of impact that AI might have than the first. I guess I don't really feel like I want to have 10x the stuff that I already have while I'm really excited about someone curing cancer, Alzheimer's, Parkinson's, MS, and so on.
It's can't without an explosion of energy use. An explosion of economic activity will create an explosion of resource extraction and pollution. Since we are already going beyond many planetary boundaries we likely risk destroying organized human life because we destroyed our ecology.
Naming LLMs as "AI" is really proving to be a colossal mistake and/or a deliberately misleading strategy that has (very successfully) helped acquire capital. Because it brings in all of the cultural tropes about superintelligent robots and gives them a veneer of respectability; e.g., The Matrix, 2001, The Terminator, etc.
Conflating the two is a massive category error. It's perfectly possible for LLMs / current-generation "AI" to have a massive economic impact without thinking there are going to be digital human replacements in the next 5 years.
"The measured economy becomes dominated by whatever it is that cannot be made more efficiently" Which is why health care support operations are the fastest growing labor category in the US.[1] And why medical care has become such a huge part of the Federal budget. Meanwhile, farming is down below 2% and manufacturing is tiny. That's because those were the areas with the biggest productivity improvements.
AI will come first for those for whom everything they do for income goes in and out over a wire.
The civilization will destroy itself first. States are not addressing what really matters and especially not the multi billion/trillion companies. The US fight against social security and less inequality, the things that wealth should enable. We are destroying jobs that brought purpose, just because some small percentages in margin. We are not training the AIs to solve climate change but to code well. Maybe it’s time for the UN to become the main government of this world, to deal well with the uneven distribution and discrimination of work force. It is too much about egos, too much thinking in cultural groups to protect from outsiders, too much war and destruction, still too much burning fossil fuels, too much destruction of our environment and the nature. Maybe with the appearance of aliens first to have a visible outer enemy first? But I guess it will be the same as always: as long as most find some distraction, some misinformation that calms down fears and anxieties, nothing will change for the common good.
I immediately think of mining. Mining in most countries places little value on miners: it is a dangerous, life shortening occupation very often done by those without much alternative. It's history is one of random deaths and maimings, quite often of large groups of people at once. It is associated with numerous health problems. Pretty much the only thing that has ever improved conditions for miners anywhere is either collective action or paternalistic, labour-sympathetic government. Mining and jobs like it are among the stereotypical situations where people imagine that education for their children will lead to a better life.
All of which is to say, will the most awful jobs actually be automated? Not just awful, either, but poorly paid? Logically it makes more sense to approach knowledge work first: it requires only one kind of investment, namely data centres. There are thousands of niche skills in manual labour that will be awkward to make robots for. I think also of old housing stock that needs to be maintained by plumbers, in particular: legacy stuff built up over decades (or where I live, centuries)
Language models seem to have come a lot further because they can use the generalisable capital of computers. Outside of controlled environments (factories, cities with grid systems) I don't imagine robots will represent such a good investment any time soon. Knowledge work is dying, time to learn a trade. Maybe.
(It would have been great to have been able to learn and practice a profession for a lifetime. Library churn is awful, I suspect what will follow for most of us here will be worse)
Here's a thought: what if descendants of today's open models end up refined by proprietary models, resulting in a world in which it is largely impossible to keep AI under private ownership? What if data protection laws give way to data donation laws, permitting people to specifically choose recipient research institutions to train new AI upon? I could easily see a world where OpenAI, Google, Anthropic etc are boycotted in favour of say, a global-membership cooperative. It'd require people to exercise some agency but few predictions about AI seem to imagine people have any. Governments are spoken about as though they're managing people rather than representing them. The likelihood of popular movements or collective endeavours seems to not have been factored in at all.
"AI", and even robotics, are terrible at construction and maintenance. Most housing-related problems are in that category. It's not at all clear how LLM-type AI can help much. There are at least 18 humanoid robots on YouTube, but we don't see them doing much useful manipulation.
That's the part that needs to scale up.
That YC startup working on robotic construction equipment is a step in the right direction. But it's mostly automatic driving.
About twenty years ago, someone hooked a backhoe up to a force feedback hand input device, so you could dig by making clawing motions with your hand. The neat thing was that you could feel your way around pipes and rocks. Never got beyond a prototype. If AI manipulation gets any good, that sort of thing should be a robot.
37 comments
[ 2.9 ms ] story [ 50.2 ms ] threadAssume those loops have maximum force and the economy becomes “information produced by information capital, which is produced by information, which in turn is producing information ever faster every year”, as William Nordhaus, a Nobel laureate in economics, wrote in a paper in 2021. This brings about the “singularity”—a point when output becomes infinite. The singularity is really a counterargument: proof that the model must, eventually, be proved wrong. But even the first step on the journey, a big acceleration in growth, would be a profound event.
Future transactions can have bots on both ends, running an economy that continually grows to service the needs of (probably more efficient) machine capitalists rather than human capitalists.
Just like when human agriculture and industrialization broke multiple organically forming food webs in the natural world, so could it be with human industry
It really is a brave new world.
When I ask people what they would like to change about their life, pretty much everyone mentions that their apartment is too small and too expensive.
Because everyone wants to live right in the center of a big city.
While at the same time, when I look at Google maps, big cities are tiny. I zoom out a bit and every big city is just a tiny speck in the middle of nowhere.
When construction becomes cheaper, it will be more compelling to build houses outside the city. And when driving in comfortable autonomous electric buses is available frequently and cheaply, living outside the city becomes more compelling.
On the one hand you can imagine that work gets supercharged, allowing companies to produce 10x the number of widgets at 1/10th the cost. The economy would grow rapidly, wealth inequality would presumably be exacerbated, jobs would be automated, we might need some version of universal basic income, and so on. People debate whether or not this kind of transition is imminent or if it'd take decades.
On the other hand, it's conceivable that not much would happen in the "bulk" of the economy while at the same time the frontier of humanity might be pushed forward. We may see new treatments for diseases, new types of energy production, and so on. In this version of the world, jobs would mostly remain unchanged (at least in the short to intermediate term), perhaps with some small multiplicative efficiency factor, the economy wouldn't grow rapidly, there wouldn't be any mass unemployment, and so on.
In my mind, I'm much more excited about the second kind of impact that AI might have than the first. I guess I don't really feel like I want to have 10x the stuff that I already have while I'm really excited about someone curing cancer, Alzheimer's, Parkinson's, MS, and so on.
Almost no growth.
Yet people are mostly happy.
Government makes mostly wise decisions.
There’s not too much poverty.
They have not devalued and destroyed their own culture.
So yeah, growth maybe isn’t all it’s cracked up to be. The price is too high. It benefits mainly politicians, corporations and the rich.
Conflating the two is a massive category error. It's perfectly possible for LLMs / current-generation "AI" to have a massive economic impact without thinking there are going to be digital human replacements in the next 5 years.
AI will come first for those for whom everything they do for income goes in and out over a wire.
[1] https://www.bls.gov/emp/tables/emp-by-major-occupational-gro...
All of which is to say, will the most awful jobs actually be automated? Not just awful, either, but poorly paid? Logically it makes more sense to approach knowledge work first: it requires only one kind of investment, namely data centres. There are thousands of niche skills in manual labour that will be awkward to make robots for. I think also of old housing stock that needs to be maintained by plumbers, in particular: legacy stuff built up over decades (or where I live, centuries)
Language models seem to have come a lot further because they can use the generalisable capital of computers. Outside of controlled environments (factories, cities with grid systems) I don't imagine robots will represent such a good investment any time soon. Knowledge work is dying, time to learn a trade. Maybe.
(It would have been great to have been able to learn and practice a profession for a lifetime. Library churn is awful, I suspect what will follow for most of us here will be worse)
That's the part that needs to scale up.
That YC startup working on robotic construction equipment is a step in the right direction. But it's mostly automatic driving.
About twenty years ago, someone hooked a backhoe up to a force feedback hand input device, so you could dig by making clawing motions with your hand. The neat thing was that you could feel your way around pipes and rocks. Never got beyond a prototype. If AI manipulation gets any good, that sort of thing should be a robot.