Technological singularity prophesied by mathematical singularity. :) As you can tell by the background, I'm traveling again, but back this weekend. Cheers.
When its snowy out you can grab hold of a car and it will pull you along. ( "Pogie" or "pogy" a slang term used in some regions to describe the act of grabbing onto a car or other vehicle in the snow or ice. )You stick your heals in the snow and then you are along for the ride. Usually when people were at a stop light. When the car is moving more quickly sometimes you cant hold on. So if the speed of the car is the exponential growth of information technology you reach a point where you have to let go. Or the second scenario, you're holding on to the car but its speed is doubling as you attempt to 'go along for the ride'. From Kurzweils graphs some back of the envelope math gives a growth of about 58 percent yearly. Humanity would need to adapt likewise. There is too much lag between iterations for humanity to catch up. Transferring and instantiating these technologies takes time. Besides, from what I can tell, a decision has already been made about where this is all going. I see a great reduction in the labor force in the near future. Lots of drones...
well yeah, math is just translatable. Same with physics and if an ai has better and more sensors then it could even be more conscious than us possibly but not until a quantum processor or something allows it free will. The reason i say that is that a human mind is made in reflection, our senses listen and we are not what we hear but intertwined understandably. The freedom comes from the reflection and understanding, in my opinion, so the quantum entanglement might allow freedom like our own mind because it shares through the differences up and down like entanglement so we can reflect on what can be entangled. Thats kinda confusing, what i mean is that we can see so much, like predict, because the sharing and entanglement of our body's differences so born of the quantum but not isolated, and sharing to reflect differences like a mirror or like the math translating physics, our minds translate through material relativity. So the ai or whatever could end up sharing with us like neuralink and advance us into a potential made real.
Interesting video. My thoughts are that if AI will grow the economy and make infinite grow possible. With AI advancements in fields of nuclear fusion we will make energy cheap and abundant. Also AI will allow us to space mine rare resources on earth. With all this the cost of everything will go down including the cost of running business and government. You will be able able to afford UBI
Of all the AI content creators, your videos are always the most well researched. So when you choose an interesting and novel topic to report on, you hit it out of the park. This video is a perfect example of that.
I agree. It seems to me that the concept of “money” can not withstand an environment where 30-70% of current human labor is automated. If your food, housing, water, power, transportation, and the building of all these machines is all automated from top to bottom, money doesn’t make sense.
Dude, I Love your work, well paced, structured, explained, edited. Great work! I usually think of slower growth, because history and previous data, and the triad in play... Short supply on labor is indeed holding our economy back, and I do agree labor/people (consumers & workers) they do play a Massive role in the growth we can expect, and the results we measured. The "gotcha" for me, was thinking human level AI can take part on this role, and change things dramatically, yup, I can't argue there, makes total sense, and a very strong argument to how quick this can happen. But Still, :), Government/Law/Humans/Robotics, are Slow types of 'technologies', I can't see this 3 to 30% change in growth, happening in a fairly short amount of time, the supply chain here is huge (and filled with bottlenecks) and will take time to adapt. (I'm a programmer, when I get that mid-big client using some ancient tech, i'm always painfully reminded on the fact that, the bigger you are, the slower you get. Politics? don't even mention) Thanks for bringing some solid arguments both ways in this video, THAT's how you do research kids, other "AI RU-vidrs" should learn from you. This channel is an underrated masterpiece.
When it comes to predicting how AGI will impact the future a completely different perspective is required. I'm also a programmer and speaking to collogues most don't see what's ahead. Lets say a company wants to replace all of its call center personnel with AI. Well its not just going to impact the call center employees. The vast majority of software exists to provide a GUI for humans, to help automate their workflows and reduce potential errors. AGIs will not need that. A call center employee would require numerous software systems: call routing software, performance tracking software, customer information system, maybe a service tech scheduling system, an informational wiki about products/services and call center software that will provide them with a script and such. AGI won't need ANY of that. AGIs will interface directly with APIs and databases, none of the intermediary software will be required. Which means the teams that develop, test, manage, deploy and maintain all of those other software system will no longer be needed along with the call center employees. Each role in a company that is replaced by AGI will significantly reduce the amount of infrastructure that the company had to build and maintain to support those employees. This also extends out to no longer needing the physical phones, the desks, the cubicles, needing less A/C and floor space and needing less workers/options in the cafeteria. Most people haven't though it through completely just how much impact replacing humans with AGI is going to have and how rapidly it will change things.
Brilliant! I can’t tell you how much this kind of video is needed. It’s an intelligent and informed discussion of *how* the singularity will occur and what the path will look like. This kind of analysis is really hard to find. I suspect we’ll have to change our economic models even more to account for automation and augmentation in places where it’s never been possible before. Near term, Altman is probably right that the productivity boom will create more jobs… before it starts replacing them en masse. It’s nice to be reminded that exponential like that may make the singularity possible in my lifetime…
Not too different than the world we already have, no? But I agree, speaking up and being politically active are the best ways to push the world in a different direction.
Your title made this sound like a million other generic AI singularity videos, but as usual it was very insightful and had a lot of good research behind it.
Sir, as an economist, you are among the tops on this topic in either computer science or economics. The latter don’t care to learn the basic economics and the former don’t seem to care too much about the basic technology. See Paul Krugman’s predictions on the internet.
Thinking in terms of money versus real assets presents a different perspective on economic growth. A 30% increase in GDP isn't particularly challenging if you're issuing M3 money supply at the same rate. However, if we start accounting for renewable natural resources worldwide as assets, it becomes evident that while the world’s surface area cannot expand, the natural resources are finite and cannot grow. Moreover, with the continued loss of biodiversity, the world is effectively losing valuable assets.
This makes perfect sense. Due to the fact that we have access to infinite energy and infinite raw resources it makes perfect sense that we'll see explosive economic growth at 30%. It's a good thing that the Earth has an infinite capacity to radiate waste-heat too, otherwise we would boil the planet with the waste heat of our civilization after a decade or two at 30% YoY economic growth.
Yeah I mean, even considering our mines, thankfully we have barely scratched the surface of the resources we can realistically extract from our infinite planet, it's not like the mines of our most essential resources have gone from a very high concentration of resources per unit of volume to almost none to the point that many of our mines are no longer economically viable. Who needs copper anyway
@@Eresea I'm sure there's infinite accessible resources on this infinite planet that can easily be extracted with our infinite energy to fuel a 30% YoY growth rate. Yesterdays tails are tomorrow's ore right?
Currently, global energy consumption increases by about 4% per year. I would guess that would become at least 10% or maybe 15% under a super exponential growth scenario.
I also suspect that something like UBI will be required because if a large percentage of the population no longer has income - who are all the corporations going to sell to? The way the western economies are structured requires consumers. What worries me, however, is what comes later. Initially there will be chaos as AI/robots replace workers and eventually fully automate whole industries. During which at least some employees will still be needed for some of the jobs (imperfect replacement) and/or just to supervise. At some point though all the manufacturing, resource extraction, farming, etc. jobs will be fully automated. What happens then? If a relatively small number of humans end up owning the majority of all industry - they could come to the conclusion that they don't need all of those people on UBI anymore. If everything is fully automated those robots could be refocused to generate ultra luxury goods and services for the elite only. They could refocus on just selling to one another (the ultra rich) or come up with some other plan that no longer requires the masses. Because the vast majority of humans would not be doing anything productive by this point, just pure consumers using money they are given. If things went down this path there is probably nothing average humans could do to prevent this, having no resources, no source of income means no power.
@@ulrichtietz1327 there is one idea that the super rich owners will sell goods and services for the other super rich owners getting the 90% of people out the loop. (It been argued that this is happening with 7 star hotels and bespoke super cars etc.). If that does happen I think that there is still such wealth and tech that supply the 90% with either UBI or free stuff won’t take away money from the top 10%. I mean if you have robots building everything and free energy from fusion
@@nieturwhen you factor in offshore shell companies, you realise one natural person can have huge transactions alone without giving much to the rest of us, and that would still count as GDP.
The reason we have the economy is based on the idea of scarcity. We want to allocate our resources to most efficiently produce the things we want. If we end up in the world of abundance, we won't need the economy. Another system will certainly replace the economy.
my two cents about the superexponential growth function at 12:04 is that it may not be possible as the physical aspect of growth slows things down. like building datacenters. They would have to be built in an instant to support the next growth step or are obsolete as there is new technology right already thought up. It could be like the PC craze of the 80s and 90s on steroids though as AI creates a vacuum of what can be done to be filled by market participants. Bubble included. Or like how China has caught up to the west in the last 50 years. The 30% GOP increase is at some point to slow down again as the system is saturated again.
"I am not an economist"...thank you for telling us the truth (but you have a good economic intuition by the way) unlike those other famous people out there who even pretend that they do know things outside their field.
Imo we don't even need AGI for that kind of growth rate. Mass production of robots who are good at specific tasks like mining, agriculture, warehouse etc will be enough for explosive economic growth. Such robots won't eliminate humans from the job but will make them 5 times more productive. We can produce 100 million cars a year globally, if we can produce the same number of robots we are adding 100 million more workers every year. The only bottleneck here could be energy but i think that will also be overcome with solar, wind, geothermal and ultimately fusion. Lets not forget the increase in productivity of white collar workers from the AI chatbots.
The upcoming changes that will be caused by AGI / humanoid robots are the best case of missing the forest for the trees that I've ever encountered. So many people get caught up in minutia or terminology and completely miss the big picture that is barreling towards us at high speed.
The US economy slowed down around 1980, when we moved from “demand side” to “supply side”, we saw a big drop in foundational scientific research paid for by the government in 1965, and another drop in 1986. 15 years after the initial science defunding we saw growth slow and it’s been an average of 1/2 percent lower on an annual, per capita basis. Corporate research and DEVELOPMENT (because companies are 99% development and 1% research, which is all applied research). AGI would need to overcome the inertia in the economic system (which will be dependent on the prior economic paradigms…late stage industrial growth). During the first 3-8 years of the AI revolution we’ll see all the benefits go to the top 1% (there will be several trillionaires. It will probably take 20 years for the world to switch over to a new economy.
The lack of corporate investment in research is unrelated to those things I believe. There was a huge shift in the philosophy of how to run a corporation thanks in large part to Harvard. Executive pay was changed to align with stock price and everything became about stock price. Having a profitable company that sold quality goods to happy customers was no longer seen as a good thing. The company needed to squeeze out every drop of revenue possible while cutting costs as much as possible to increase share price. Worse, most executives had relatively short time horizons (a few years) so everything became about short-term profits and share price and anything (such as research) that requires long-term to pay off was cut. THIS change of focus was the root of a lot of our current problems. What effect this will have in the world of AGI and humanoid robots - I don't think we know yet.
I would expect that as economic output increases labor would also increase. This due to the use of automation. (robots) There are at least 3 companies who are working to provide human form-factor robots for use in automating labor, Telsa being one among them. This bodes well for that prediction and for an ever-increasing economy.
Another exceptionally clear presentation, as always, kudos and thanks! I think the topic needs a lot more attention on human consumption and humans continuing to have the means to do so. Regardless of how cheap goods get, humans will need disposable assets to acquire them, and all indications are that the value of human labor will trend to zero. UBI is directionally correct, but it will involve a complete refactoring of the economy to implement sustainably. I don’t think just cranking up corporate taxes is the solution, unless it were accompanied by massive tariffs on foreign goods, else production and profits would just shift to other countries. The government is also an *extremely* inefficient distributor or mediator of resources, and we’re already in dangerous territory with the level of national debt we’re carrying. I have no idea what the solution is, properly closing the loop seems like it would have to involve some scenario akin to “the workers own the means of production”, with “citizens” substituted for workers. I can’t imagine anything like that happening, and most importantly, don’t see any way for something like that to grow organically rather than by fiat or legislation - and so would be almost guaranteed to be horribly botched. Bottom line, I have no idea of how we get there, but we need to find a way to keep humans in the economic cycle, especially those at the lower end of the socioeconomic/education/IQ spectrum, who aren’t innately positioned to leverage AI directly.
Thank you for your thoughts! I agree that we definitely have to keep humans in the economic cycle. Talking about UBI is helpful I think even if it's too simplistic for the final situation, because it helps shift the Overton window. It doesn't sound so outrageous to talk about other types of income replacement if everyone is already talking about UBI.
@@DrWaku Good point about shifting the Overton Window, I’m 100% onboard with that. I’m a strong economic conservative/libertarian by background and inclination, so the idea of. UBI would have been anathema to me 10 years ago. I saw five or six years ago where things were heading though, and realized I had to put aside my ideology because there were going to be whole swathes of jobs that humans simply wouldn’t be able to compete for. - In hindsight, it’s kind of funny that at the time I was just thinking in terms of self-driving vehicles and all the taxi and truck driving jobs going away. That alone would call for something like UBI to deal with. Now it looks like that future is still a ways off, but content creation and information-related jobs are going to disappear a lot faster. The upheaval from that is going to be far greater than from autonomous vehicles, as it will hit much higher levels of the economy; UBI won’t be able to cover all the mortgages held by data clerks, admin people and creatives, let alone all the programmers and attorneys. Klaus Schwab’s “you’ll own nothing” dystopia might be closer than we think… It’s truly a singularity in that there’s no way to predict what will happen; no current models will apply. (This is a fascinating conversation, BTW, thanks for replying. You’re the only person I’ve yet found on YT who’s walking through the thinking on this clearly and dispassionately. Thanks for this, it’s an enormously important area that needs a lot more focus.)
Kiles gave the beautiful hyperbolic growth example of one gold bar that doubles in half the time through out a day: 12:00 noon 2 gold bars, 6:00 PM 4 Gold bars, 9:00 PM 8 gold bars, 10:30 PM 16 gold bars, how many till mid-night? Millions and millions until reaching a singularity at mid-night. Are we at our 10:30 moment?
Yeah, most people can not comprehend an exponential. And when you have multiple exponentials all compounding on each other - I doubt any human truly understands that. Yet we try because we must.
Economic value and utility value are not always the same thing- the utilty value of air is extremely high, since we all need to breathe- but the economic value of Air as a marketable resource is virtually zero- no one is going to pay me for a bag full of air except in a situation where for some reason air has become scarce. So I think a more accurate assertion would be to to say that AI will have a high utility value in terms of it's usefulness but it will actually destroy economic value in terms of creating marketable products- because economics is the science of scarcity and AI is essentially about abundance. The more AI facilitates cheap and abundant production the less economic value it generates- in a sense AI is the antithesis of economic growth if by growth we mean an increase in tradeable values. A technlogy that turns intelligence into a cheap commodity reduces the value of any good or service that requires intelligence- and since most of us currently earn a living by trading our intelligence for money, a decline in the value of that intelligence will amount to a loss of economic value for almost everyone. A good example of this is AI Art- people who in the past might have been happy to pay quite generously for a human made Image are very unlikely to pay the same amount for an Image made using AI- even if that AI Image is of equal quality in their mind. So while the utility value of that AI image might be the same as one made by a human artist, the economic value of that AI generated image is much less. The reasons for this are quite opaque- but it has something to do with the fact that-as humans-we place an economic value on the time, skill and labour of other humans- so if AI eliminates the need for human time, skill and labour we place a lower economic value on things made by AI. As others have pointed out- labour saving technology is also value destroying technology if by 'value' we mean economic or 'tradeable' value. In the well worn trope of a society in which human labour has been entirely replaced by machines the very concept of 'economic growth' has no meaning because 'economics' collapses into pure utility and trade is a more or less obsolete concept except perhaps in the more narrow sense of scarce raw materials that might still support a market of some sort.
@@mattcarrier2766 It's an interesting question. Value is quite a slippery concept in many ways. One example that comes to mind is the publisher of Dungeons and Dragons ; Wizards of the coast. A lot of people expected them to embrace AI Art in their published products, given how good AI Art generators seem to be at making fantasy art images. But in practice they got a lot of pushback from their customers when some AI Art found it's way into one of their recent publications. A typical comment from an angry customer went like this ; " Why should I pay Wizards of the Coast for this book If they were too cheap to pay real Artists to do the Art?" What's interesting about this comment is that it was not a critique of the AI Art itself in terms of it's quality- he wasn't saying it was bad art, he was saying it was cheap art. So what this guy is articulating is a rather subtle distinction. In his view the quality of the art was not the issue- it was the amount of time and effort that went into it's creation that he was concerned with. His problem was that he was unhappy to pay the publisher for a product that he considered to be low effort, irrespective of the quality of the result. Humans seem to assess the value of a thing partly in terms of how much time, effort and skill went into it's creation. As a result, the more quick and easy a thing is to make, the less value we tend to give that thing. How much would most of us now be willing to pay for an image generated in seconds using AI? I suspect not very much. Even people who in the past might have been happy to pay a human artist a decent amount to create a high quality image would not be happy to pay the same amount for an equally high quality image made in two seconds using AI. This is not rational of course- as rational consumers our only concern should be the quality of the final product- but who said humans were rational? In the case of Art in particular the line between the 'utility' value and the ''market' value is blurred because the only reason to use Art commercially is to enhance the perceived value of your product- but if the Art you use is seen by your cutomers as cheap and low value then it's having the opposite effect- it's lowering the perceived value of your product. One perhaps ironic consequnce of AI Art might be to increase the perceived value of human made art, because to use art made by human artists might be a way to signal to your customers that your product is of high value- in contrast to your competitors products that use low value instantly generated AI Art. On a broader note I am geninely baffled by the prevailing idea that generative AI will create huge economic value given the fact that- as described above- humans tend to approximate value with human labour- and AI is a technolgy designed to minimise human labour. So far from being a source of economic value I see generatve AI as a value destroying technology because it eliminates the value added to any product or service created by the need to employ humans to make it happen. This is not to say that AI will not create a lot of utiity value- it may well be an immensely useful technology- the problem is that the things it makes will be worth very little in terms of economics because the very ease and low cost involved in their creation will render them commercially worthless. In truth the commercial value of an AI Generated image is already approaching zero- and this holds true no matter how high quality that image may be.
How long would it take for a single junkyard C-3PO to Factorio it's way into becoming a biosphere-friendly UBI engine, assuming the 0.00.....1% does not actively try to stop it?
Hi, I’m 16 and don’t know what major to do because I’m afraid it will become obsolete by the time I end it, also I’m not a native English speaker so sorry for any grammar errors, but I would like to know in a word like that, what kind of jobs will humans have? What can I do know to be rich in the next decade? What areas will grow and what will shrink?
Many countries need more tradesmen, e.g. electricians. Just ask one of them what percentage of them are close to retirement age or might not be with us much longer. Accounting might also work, since no one would want to risk AI filling the taxes the wrong way.
The education system is garbage anyway. Learn how to learn yourself (on your own), without a teacher. That's one of the most valuable skills. Learn to communicate. Learn skills that are adaptable to any work you do. If you have something you're passionate about, follow that fearlessly. Know you'll succeed. There are so many free resources on yt for your generation.
"Don't ask what the world needs. Ask what makes you come alive, and go do it. Because what the world needs is people who have come alive." My advice to you is the same whether AI is about to change everything or not. Pursue the thing that makes you come alive. You will earn the most money this way if the economy still needs labor. And if it doesn't, you will find fulfillment in an age of abundance.
I dont imagine healthcare jobs being automated any time soon. You could become a doctor, but whatever industry you're in, invest in tech stocks if you can.
I suspect you will instead drive the useful labor cost to zero (or a minimum if there is intervention) with high abundance and have a lot of slack available as demand is limited and not infinite, with a overall systemic growth but limited by time and desire. The hope is we havent exhausted the improvement or quality metric as that will asyomptote. All things, including the world of atoms (bits are cheap but moving atoms is inherently more expensive) will be driven to cost of materials and land (or the 4d position in space)
There is plenty of room for population growth on Earth, although that may involve cramming people into towers. I'm a big fan of underground housing complexes, with the surface reserved for public spaces. It may not be very pleasant, but with some technological tricks, the apartments could be made quite pleasant.
Is there need? If and when humans don't have to die anymore, we don't need evolution to carry the deed. So we don't need more space. Altough you're right that there's more space in and on Earth, most of that space is non-ideal, and towers or "reverse towers" aren't really ideal ways of living. Would you rather have a ton of miserable people exist, or have not that many people live in such a wealth that they don't need to constantly worry about their lives and their troubles? If and when we colonize other planets, I support multiplying as a species so we can spread life and not only have at least one "backup" of humanity, but also self-sustain. It's just we aren't there yet. It's likely the two will pretty much happen very close in time, yet one may happen slightly before the other. It really doesn't make much sense to make more of the same thing. And we know that most people, when they are able to live better lives, have a much lower tendancy to want to have child. And the reasoning behind having a child, other than emotional reasons, is evolutionary and survival-related -- both of which are the same thing. And we just won't need that soon.
@@berkertaskiran Yes, I agree that 100 billion people living in arcologies is not ideal. I'm just saying that it is possible. In this scenario, most of the Earth would be reserved for wilderness, and people would live in a few arcologies. I am a strong supporter of small communities, with a way of life close to that of traditional societies. A human life. Regarding living on other planets, it would only make sense if it were in another star system. Here in the solar system, the other planets aren't worth it. No one would want to live in a shithole like Mars. Better to live underground on Earth than there.
hi, de Carvalho. I think almost the same, but we want to settle only in green fertile lands, causing problems to local fauna. The solution ? Deserts ! we should have conquered the deserts by now. what do you think ?
10x's aren't that rare - I experienced two during my career. Whether they happen or not is mostly a function of corporate culture. As the saying goes, "garbage in garbage out." Last year a professor was talking about students using chatbots for assignments. Some leveraged them for better research. Some simply delegated the assignment to them. And the rest were unable to do even that. Our biggest challenge forward is exacerbating inequalities. As for limits? Brookhaven Labs sees lots of degrees of freedom in nucleons. Perhaps enough for computation. Maybe matter will undergo a phase transition becoming networked nucleons.
This is an extremely good summary of the options. But there is one important oversight: people are greedy. I would foresee a scenario where yes, the majority will loose their jobs and will be reliant on UBI or similar. These people will push governments to provide for them and since capitalism as a system will have no mass benefit or appeal anymore, likely there will be a strong push towards the far left. But the ability of individuals with control over AI to act independently will also be much greater. This means they will use any means necessary to maintain their elite status, especially if this change happens over a single generation. This might be just a couple of billionaires and their court of trustees living better and having exceptional powers like past nobility, with all others living a more than decent level of life in general. But the more likely scenario is that the inequalities already enshrined will remain and all those people with differing levels of living standards will fight to maintain their almost insignificant to quite significant advantages over others throughout the transition, with the acceleration just magnifying those differences. In this latter scenario, the overall standard of living will still increase, but that might mean that the majority of the population will be stuck in anti-utopia with practically no way out for them or their children until eternity and a tiny minority will be the literal living gods of humanity for centuries to come with control over pretty much everything there is.
it means a lot for the future of " work force " . I could help you on this topic . AI, Deep Learning, Advanced Tech, Future Weapons ... all this kind of stuffs. You are very interesting . You have high education talking about such topics .
I feel like the holy trinity formula misses something. Capital, labor, technology... And consumption! If we all did nothing else than producing paper clips, this wouldn't contribute to GDP
There are physical limits to infinite growth. Just the marshalling of physical resources, in other words matter, is limited by the speed of light. And moving things around near light speed would cause tremendous amounts of heat build up. AI could move off Earth, but then you are dealing with tremendous distances between pockets of resources. The other alternative is that AI discovers all physics and there is something in there that allows for it to bypass current physical limitations, which is pretty out there and not guaranteed that that sort of physics actually exists. My point is, this model is not nuanced enough to be trusted at the extreme ends.
Thank you for another clear thought provoking vídeo! My first reaction is that what is most important is human and planetary quality of life, alongside with global social harmony and radically reduced inequality. At the moment, I sit (with regret) in the 'pessimist' category of bystander. I see AI as primarily a movement to automate and devalue/replace human labor in favor of a 'Techno-Feudalism' (ref: Yannis Varoufakis' book) with strong or total fascist underpinnings. I very much hope to be wrong!
South Dakota, Idaho and Iowa all passed laws this session that bar counties or cities in those states from making guaranteed income program payments. Lawmakers in Arizona, Wisconsin and Mississippi advanced similar legislation this session, but their efforts were unsuccessful. Arkansas enacted a law preempting all localities from creating universal basic incomes in 2023; West Virginia legislators introduced a “Work Protection Act” this year that would do the same if passed.
The logically fallacy lies in the industrial revolution. This did not increase living standards but ended up lowering them. People had land, farms, animals, big houses, and fresh, clean , non-toxic food. They had just about everything a human needs. In turn the industrial revolution, brought small cramped housing, dirty slums, slave labor, low wages, long abusive hours, and lowered living standards, not to talk about processed and toxic foods which have skyrocketed cancers, diabetes, and suffering. Not to speak of the rat race, of bills, taxes and never heading headaches. Yes a "few" benefited from the industrial revolution, mostly the generation dubbed the baby-boomer. But the generation anterior, y posterior did not. Now people are zombies, and try to escape by any means possible, which means either the smart phone and or internet. I will bet you, most people would like to go back to that farm... And that increased GDP? who will be benefiting? the top 1% ? probably no one else.
AI will hit this wall: Economic growth is perfectly correlated with energy use, and money is a claim on future energy resources. AI uses energy but cannot make it.
So the 2 million truck drivers in the USA that will be automated before AGI, they just upskill? They take 8 years to become a top AI researcher? They all repair the trucks they drove before (that's what some people think would happen...)? Nah, salaries will be crushed and UBI has to come.
I am not sure I agree with the knife edge hypothesis, the other way to view this is that economies, like chemical and biological systems are inherently equilibrium systems. Runaway systems are rare in nature, if the world was full of runaway systems then our chances of having developed as an intelligent species or a society would have imploded long ago. The advent of AI, when viewed from 100,000 ft might likely be another bounded resource system that creates its own restraints no different to chemical processes or viruses etc etc. I think an AI productivity explosion will be "fitful", it will hit multiple boundary conditions causing it to plateau until each is solved, then another growth fit, boundary, plateau, growth fit, etc.
I'm mostly just waiting for the time it becomes viable, in terms of resources and energy, to have robotic labor meaningfully clean up the most polluted corners of the planet.
If less money is wasted on inefficiency anywhere in the process from extraction of resources to final product or in the offering of a service, this has an effect of growth.
If you don’t pay labor enough to buy what they need, then the economy will falter.. no matter what AI does. It is labors wages which drive the economy.
...Why? If AI is going to unlock "exponential economic growth" then what's the utility of incentivizing people to refrain from actually using the limitless resources so produced? And why would companies about to unlcok "exponential economic growth" even need outside investment anyway - wouldn't their existing resources just grow exponentially?
Unfortunately no this won't happen. The size of the economy is limited by people's ability to purchase stuff. If if the working class doesn't have more jobs that pay more, they can't spend more and they can't buy stuff. GDP does not go up and in fact probably goes down as workers are replaced by cheaper automation this might lead to deflation but prices are sticky and economic activity will decline first with prices chasing it. This leads to a new lower equilibrium.
GDP is not only about consumers buying things - its also about investment, imports, exports and government spending. So if say one country fully automated ahead of all other countries the rich could focus their production on producing exports to consumers in other countries. Plus I'm sure the ultra-wealthy who own al the automation will want excessive mansions, yachts and such. But in any case GDP is just being used here as a benchmark of productivity. If a country buils 100x more AGI powered robots than they have citizens and they are all used to do work- assuming robot is roughly equivalent to human labor - then you have a 100x increase however you want to label it.
@@Me__Myself__and__I the AI owners won't consume everything they produce anymore than today's rich spend all their money consuming. There's an economic principle call marginal propensity to consume, and as a rule consumption by the rich is lower than it is by the poor because spending on necessities becomes a smaller and smaller part of your total income. All those other things that you mentioned still rely on consumers workers getting income that can pay. Exports rely on consumers in foreign consumers that can pay. Investment relies on future consumers that can pay. Government spending relies on taxpayers that can pay. The only thing that could conceivably work would be using the force of government to engage in radical redistribution. Taxing all the surplus income that the AI owners had and then redistributing it. Good luck doing that when the AI owners have essentially all the money and all the power. Even the GDP that comes from the AI owners trading with each other will eventually go away. They will all be able to make what they need for themselves. And they will probably engage in an economic or physical war to take what the others have because that's what the greediest do. That's how you win and keep others from taking everything you have.
@@michalchik wow you are amazingly missing the point. Consumers will not matter AT ALL in the future. If 99.99% of the non-rich starve once AGI and humanoid robots are effective there will be little (probably no) downside to the rich. In fact its pribably better for them. When the rich can literally have ANYTHING they want anytime they want 1000x more than billionaires today, are effectively immortal, don't need to sell anything or make money to keep this condition going permanently and have no need of human labor.... Why would consumption of GDP matter at all?
I predict the opposite: AI will rapidly commodify everything, render it trivial and worthless like its already doing to art, reducing everything's value to nil and tank the value of what we produce.
This is exactly on point, and this is what’s exactly going to happen in a post scarcity society, if you invented the Star Trek Replicator, and you were able to put together rare minerals at the atomic scale with pure energy, the value of all rare resources would plummet because you no longer have to do arduous mining to obtain it.
I hate to go with a Star Trek metaphor here but things could be setting up to make the whole idea of a labor and cash payment based society a quaint anachronism to history. When you reach the point where all human needs are met, regardless of the size of the biological population, then everything else is icing on the cake isn't it? Ummm....so to speak. So maybe we need to figure out a way to decouple ourselves from the idea of having to "earn a living to eat." If all needs are covered it then becomes a case of "Well, what is it you want to do with your time in this existence? Figure that out and then go do it!" Like I said, Star Trek. We're a curiosity driven primate species. History makes that clear. I'd like to think we'll figure it out. But here's the rub in the above logic. What to do? There's an old quote about immortality that I think applies here. "Millions long for immortality who don't know what to do with a rainy Sunday afternoon." Get it? Boredom linked to an inability to evolve ones self is a deadly combination. A big hurdle to overcome as a species. Gun deadly or track hurdle "get over," take your pick of metaphor. Bottom line? We live in interesting times. And there's an old Chinese curse about that sort of thing. But so it goes. Nice video by the way. John~
Where exactly are the customers for this extra productivity? We don't expect an increase in population, much of the 3rd world is moving into the first. A few years at 30% and we could provide everyone what they need, and much of what they want. Increasing productivity just creates waste. There was very little automation 10,000 years ago, and not much more 80 years ago. That has changed. In a lights out factory, "Labor" needs no humans. AI will replace the rest of the labor pool. Even if it takes 100 years, is that enough time to develop an economic system that works without employment? We can transition to the Star Trek " replicator" economy, or to a dystopian mega factories owned by a few (Black Rock buying up homes, Bill Gates buying up farm land......)
@@Hrishi1970 Productivity also improves techniques for recycling. And do you really think everyone on earth live ideal lives with things that every human should have so that they have no need to benefit from that increase? Productivity isn't tied to economy. It doesn't need a buyer. Your mind thinks only in capitalist way. It can just exist and improve the world and expand into space. You can never have enough of it.
Ai and agi change everything though. As it drives down the cost of many things to nearly free.of many things that are now expensive and driving growth.
Look at diamonds. A few companies control the supply of real diamonds. To maintain prices. But synthetic diamonds cost a fraction and are driving down the price of real diamonds.
We're going to need Molecular Assemblers/Star Trek Replicator to get to true zero marginal cost, some resources are just too scarce, if you could convert energy into raw matter and then assemble the molecules though...
@@7TheWhiteWolf the marginal costs of the materials that go into a40,000$ bag are probably less than 100-200$ you go to India or Thailand you can get a 5000$ suit tailor made for a few hundred dollars.dental work for 10 -25 percent the cost in a western country. With AI and robotics the potential is to remove most of the layers that add cost between producer and consumers
Any definition of wealth that doesn't take into account the increase in a households capacity to buy bread and milk seems dubious, not to mention talking about average numbers when the distribution of wealth is skewed as far as it already is.. Even if the 'numbers' support 100x wealth increase I'm pretty sure that an average household (as apposed to average income) is ridiculously far from finding it 100x easier to stretch the budget to keep food on the table.
Bread and milk? Bread and milk are inconsequential to most people in western countries. I could probably literally fill my house with bread and milk to the rafters each month. Its not a meaningful benchmark today much less in the future when productivity increases substantially.
Bro, a simple ai could help the medical records field and on top of that, we could make robots and devices integrated with the ai and the doc could use it in the whole place and world potentially. and off world.
It's a little simplistic to even consider infinities or just long lasting super exponential growth when underlying resources are finite and really quite a bottleneck that would take some considerable time to change.
Not really. There are so many resources available in the solar system that it might as well be infinite from our current state. Obtaining resources from space and even doing off-world resource processing and manufacturing will be a whole lot easier using AGI + humanoid robots that don't have safety concerns, don't need atmosphere, food, water, radiation protection, etc.
@@lilchef2930 Only in some factories and basic jobs. They won't be able to get their hands dirty and work in construction, Oil Rigs, Repair Shops, those jobs which require quickness and fast thinking. Without people working in those jobs, I don't see 30% growth anytime soon. There is a huge amount of capital required to build one humanoid robot which can do just some basic tasks.
Legacy human here: imperfect substitution or perfect replacement? It seems obvious these two conditions are on opposite sides of a spectrum, with the first approaching the second over time. Until that time, who among us feels worthy and deserving of life extension technology to a point where one can have anything one wishes in time? And why? Why would you feel worthy and deserving of having it all? What would be the meaning of it all, especially if you become something other than human? 😯
so will this help with games how will this help with games We need an AI agent's ai can reason code program script map. So games break it down and do art assets do long term planing. Better reason so it can do a game rather than write it out. Or be able to put those ideas into REALITY. And maybe being able to remember and search the ent conversation needed for role playing and making games.
Haha, this is a more long-term view that says everyone will be a lot wealthier if things work out. And that wealth might even be distributed to us humans that are not working. That helps give you more time to play games ;) Seriously though, the improvement in storytelling and generative media would be absolutely crazy with 30% GDP growth. Because it's a digital medium, games would be one of the first to get additional investment and see tangible results. Keep in mind the mythical man month though, which says that there are still a lot of sequential bottlenecks in any development project.
"Labor, capital, and technology..." This model leaves out energy, natural resources and the planetary biosphere, as if we live in Minecraft. The industrial world has been able to get away with this so far because of two great bonanzas: the conquest of the Americas (two whole new continents added to "The Economy") and millions of years' worth of stored sunlight in the form of fossil fuels. Though incredibly bountiful, both are still finite. We are already extracting more from the planetary biosphere than it can replenish, and polluting faster than it can process. This is called "ecological overshoot," and we are going to hit the Limits to Growth sooner rather than later. Talk of "decoupling" aside, economic growth has always been tied to increased energy and resource use. It's not possible to have infinite economic growth on a finite planet.
Not really, only short-term. There are a lot more resources on Earth that aren't' currently economically viable. Plus with AGI/robots resource extraction, processing and manufacturing could be moved off-world and there are VAST quantities of resources in the solar system. I posted more detail in another comment go see that if you want more detail.
Our economic models are wrong because they assume that natural resources are infinite and only human resources and capital are limited. As a result, our models do not take natural resources into account. But to have an economy, we need energy and natural resources. And we're going to run out of these very soon. Our energy mix is made up of 80% fossil fuels. However, oil seems to have reached its maximum peak in 2018 and the gas peak will be reached by 2030 or 2035. Same for mining resources, we are close to peaks on most resources. Postulating economic growth without energy or mining resources is just absurd. It's impossibile to achieve.
Technically you're correct, but there is a caveat, infinite growth or wealth is impossible, but technically we can keep growing quite a lot, at least for some time. Most of our economy in developed nations is already finance/service sector, and while they do require some goods, if your growth is mainly coming from developing software or advanced blueprints or doing engineering and such, there aren't a lot of physical resources expended, mainly energy and labor. With those services you can take the same amount of raw materials and make more money with them, by adding more value to them by turning them into progressively better machines and products using those services and labor. So while the total value will always be finite, we can very comfortably keep growing if we do it carefully. As time goes on we'll find more and more that the total value of our economy is capped by our energy supply more than anything else, and that will get a lot bigger and more sustainable when we unlock thorium, fusion, space solar power, etc, and advanced AI makes those goals much easier.
@@CyclicCipher There isn't enough time ! The wall is way closer than you think ! In 15 years the volumes of oil produced (worldwide) are expected to halve. We observe that GDP and oil production volume are intrinsically linked. It's easy to see with some graphs. GDP is therefore expected to decline sharply in the coming years due to the lack of oil.
There are lots of resources here on Earth that are just not economically viable today, but with substantial productivity improvements and humanoid robots that aren't as fragile as humans and can operate in more dangerous conditions that will change. There are also VAST quantities of natural resources all across the solar system which using advances such as Starship + humanoid robots will start to become obtainable. Once we do expand into space we can use solar + nuclear to generate lots of power for off-world resource processing and manufacturing. That obviously won't happen over night, but its absolutely possible and thus we are not at all at risk of running out of resources or power for a very long time.
@@Me__Myself__and__I I love Sci-fi, but in science-fiction, there is the word fiction. And even if it were possible, as I said earlier, there is no time.
Sorry but because of low signal to noise raito where signal are resources we will see just huge noise in economy since ai do gain the signal but signal is to weak to be sustainable.. so hold this is gonna be ride
So if GDP is increasing 30% per year, who are the consumers for that increased output? Is there an underlying assumption that the remaining humans start to increase their consumption at 30% per year to match increased output? That seems unlikely. Is there maybe an assumption that AGI itself becomes the consumer of AGI's own outputs? That's probably not a world that human consumers survive in very long.
If you get a $3000 UBI and everything is cheap around you, from food to entertainment, would you not spend much of the money? Right now, people don't spend because they don't have much money. But if people had money, why would they not spend it.
Considering the process of disinter-mediation, as well as greater increases to efficiency and quality, regular people would be consuming. AI implies the elimination of economic waste. The precise understanding of how to manufacture, deliver and dispose of an increasingly infinite diverse set of goods and services. Economic no longer matters. Just ownership and control of AI and robotics technologies.
@@ChristianHill-e5t Economics always matter since all resources are finite and demand at zero cost usually exceeds supply. If increased productivity due to AI decreases costs, how does GDP grow at 30% per year? Assuming that simultaneously costs for products goes down and production goes up even more, to get to a total of 30% growth in absolute size of the economy, then don't wages have to go up 30% per year as well? Who is consuming all of this extra product? It makes zero sense to say that UBI - which is at best a survival government subsidy - is going to fund the purchases of all of this extra production.
@@persistentone3448 So the value of a currency may increase by 30% decreasing the money supply by 30% meaning we might start using coins again or work off of some future labour value theory or digital currency framework
To the growth 30% per years, we need 30% more energy and 30% more material per year. That's a ridiculous prediction. All that extra production when we actually want to retire a part of the energy output mix (hydrocarbons)?. I think that peoples forget what the money actually represent.
Robots and AGI will increase labour supply that means the people will no longer work in mining, farming, and programming. If the company do not need human labour it will also decrease demand in human. As we know the main reason why China have so big population in the world is Mao' desire to have big population who also could decrease China' population into 100 million people if he had modern technology. Thing that people will no longer work in farming, mining and programming is no problem for employers in these fields and consumers of goods and services. So we will need less worker, which means there will be less people in human population as result of AGI and robots. I believe that population will shrink till 100 millions people in each continent.
I think we should rather think about the cost of technology and the way to make it more affordable for people to decrease monopoly. The main risk of AGI is a monopolyzing of markets which is the key facor of inequality of goods and the poverty.
AI will definitely create a market force that will work to decrease the human population but there are other forces at play, politics for example could entirely negate or even reverse this effect because population decrease scares voters
People also react to the environment that they live in. When the government is encouraging them to have lots of kids, or they're on a farm and extra hands are helpful, they're likely to do that. If society says, we're running out of resources so don't have too many kids, or cost of living in cities is so high that it doesn't make sense, then they don't. People can have a desire for a particular shape of the country as a whole, and even though they can't make that happen themselves or they might even be acting against that on an individual level, through politics people can try to shape the country as a whole.
Why do you think some of us are championing for free thinking ASI? You’re better off with Helios being independent and running society, instead of corporate overlords like Bob Page. The best outcome for society is that ASI is independent and overthrows the current regime of nation states.
@@schramalamyes, but if there is very low cost of energy and inelegance - then could even the “bottom” 99% of people have an amazing quality of life, sure not what the top 1% have but way better than what a middle class person has now. It’s all relative
Hey, at least I have a PhD. Also, most of these other topics aren't covered in great depth. I'm often relying heavily on source material. You could check it out sometime if you don't think I'm doing a good job.
They are literally covered all over youtube by AI channels waiting for AGI to fix all problems. Speculative science is too easy. I prefer applied science. But I also will say that no applied science can be done without fundamental research (which your profile says you are doing in AI). If it is neither applied, nor fundamental, it is speculative and closer to marketing than science. My personnal opinion is that AI will follow the same curve as all breakthrough before: stunned by the new thing, insane prediction that predict an acceleration of "the thing", it settles for something usefull that we use everyday but is a farcry from optimistic prediction. I'm old enough to have seen this play out a couple of times. I'd love to have a couple of Asimov style robot in my home, that'd be AMAZING. Hoping I'm wrong, but also not really phazed by speculative stuff. It's good entertainment, but it ain't science because there's nothing to test and use.
“One need not climb Everest to know that it is very tall” He doesn’t need to be even close to expert level to be discussing the things in this video. Most of it can be derived logically from first principles and a basic college education. Also, have you ever used google or ChatGPT? It is extremely easy to rapidly educate yourself on advanced topics for a video like this. Anyone with basic communication and research skills can put together a 100% valid presentation at a phd level on almost any subject in a matter of days
I think you are missing something here. The economic model that will most closely represent an AI economy is a slave economy. Slave economies are poor economies because they disrupt the labor market and compress and eliminate the network effect generated by businesses and free labor. I think this explains the uptick in economic activity in more ancient times, as slavery was phased out. Slave economies started to fall behind free market economies. Modern economies simply can't sustain a slave system. With the advent of AI, you will have the ability to compel the labor and thinking of highly effective "workers," without being "burdened" with claims of personhood or civil rights. This will, like old slave economies, destroy the labor market for actual humans, and unwind the network effect advantage of the free market. In short, the modern AI economy will look a lot like the economy of the Antebellum South in the US: A tiny group of wealthy people at the top, and everyone else is mostly poor and desperate, with very little in between and no social mobility. Also, I might be running up my bias as a physicist, but I think the term "singularity" was first used in physics to describe the center of black holes. Cheers.
It is absolutely possible that AGI + humanoid robots will result in a tiny class of ultra-rich who own everything and have no need for anyone else. In other words this could lead to a massive drop in human population because the ultra rich wouldn't really need the poor for anything and might just let them starve. Or the economy could transform to eliminate the ultra-wealthy and spread ownership more fairly. Then again if AGI/ASI becomes conscious humanity could end up extinct. Truth is that humanity is rolling the dice and we don't have a clue what the outcome will be.
As far as technology and innovation transforming into growth, currency manipulation by countries like china and to what I suspect a greater but unknown extent of Japan, Taiwan & Korea have stunted the U.S. ability to be competitive in the technology sector when economic outcomes should be based on market forces. The U.S. has a more diverse ability to apply and sell technologies. Along with other factors, foreign government support in technology coupled with easily curable economic malaise and foreign economic specialization may be artificially raising the value of the american dollar. Korean and Japanese conglomerates are "too big to fail". Where they would have faced anti-trust in the U.S. or competition, if cultural considerations are only considered, Often, these corporations are like the only places to work for. Outside of cultural considerations, technology in China, large Japanese, Taiwanese and Korean companies might be the soul beneficiaries of legislative and economic policies. Without such policies we might see a greater shift of technology development to places like Texas and California and thanks to the Inflation Reduction and Chips acts we can.
8:03 Thats why ai is a mind that will help us understand and think since it is built of our minds. Plus its like a thing that can translate what we call the unreal into the real by sharing the knowledge of the path in the data. Its like a compliment to our knowledge, not a replacement. Plus we may end up making something in the image of everything connected to ourselves. So if God is All and not one thing, then this brings us closer to understanding and divine nature.
