Great article! Particularly about the future roles of human labor due to our comparative advantages with respect to AI.
I believe another important job for humans to perform is providing responsibility for decisions, including culpability for bad decisions. Eg, AI can already answer many common medical questions with comparable accuracy to human doctors. [1] Yet patients will demand human responsibility for major medical interventions, and even for the decision to do nothing. That ensures that there is someone to fire, sue or jail for egregiously bad decisions.
In looking at the future of work, I think we can take inspiration from the role of Vice President at financial firms; a relatively junior role that can make up the bulk of the employment for many key divisions. It’s my understanding that this is not just title inflation for the sake of workers and customers but instead a legal necessity. Many actions taken on behalf of a bank can only be performed by an executive.
Even when a firm itself can be culpable, and thereby sued and fined, we humans simply demand a specific person take responsibility for important decisions. With AI bringing massive amounts of information and intelligence to task, our society can now make many more decisions with higher quality. Yet we'll still want a human to blame when things go wrong.
This is by far the most dangerous invention of mankind, there is no way this will work out well for humanity. Look at what has happened to our society already due to unrestricted and unregulated growth and utilization of technology. Once again this will funnel all the money and power to a few technocrats and all of humanity will be their slaves, just like now only worse. The internet should have been subjugated to common decency standards like the rest of the worlds media and chat ai should never be made publicly available. Is it reasonable when each of us has access to something that can easily wipe humanity of the face of the earth, I don’t think so, it’s like giving everyone the nuke launch codes and waiting for me to launch them, if they have a bad day it could happen. The only way to allow this technology safely is to legislate strong legal and tax obligations on all robotics and technology to protect society from what has already happening. For every machine that displaces a human the owner of that machine should be required to pay the equivalent tax to cover the cost for supporting the human. Social security, unemployment, medical care and retraining if possible should all be the the responsibility of who ever disrupted the system by elimination of jobs. Humans must have jobs to function normally, only a very few can stand to not do something. This should be the first priority, keep humanity engaged in life is the crux of life. If people are unable to see a way to the future they will give up and that’s what’s happening now. It’s hard to see how this technology will benefit mankind for more than just a handful of the wealthiest citizens who don’t need any more of anything least of which is control over the rest of us.
I don't see how humans could ever have a comparative advantage with AI in a world where AI can do every task better than humans. You can use ChatGPT to write articles for more than 99.9% less than a human writer. If ChatGPT was better than human writers, there wouldn't be any humans willing to work for 99.9% less than minimum wage to try to compete on price. In any industry where AI has an absolute advantage & it's legal to replace humans with AI, human labor is worth nothing. It's certainly possible that we AI never gains an absolute advantage in certain industries. Childcare and elderly care seem like industries where people may always prefer human labor even if they cost hundreds of times more than robots, but that means humans have an absolute advantage in those industries.
So this guy's bias is basically 'concern for the economy and investing', he's really just asking the question "will global GDP fall as a result of AI-driven unemployment", and his thesis is that "no it wont because humans will continue to want more things and have higher expectations" and so the "services economy will rise". Like, assuming human services will still be considered important in the world where an AGI is making all the decisions, and human consumption will continue to be affordable (it wont be if mass numbers of people are quickly redundant). Also kinda missing the point that physical work is the first thing we've tried to automate because no one wants to do it. Great, so our software engineering, financial analyst, doctor, and lawyer jobs will go, but there will be plenty of jobs for us in delivering packages to peoples front doors, and having conversations with other humans about the surgery the AI recommends to patients. So we dont need knowledge anymore, we just need to be able to do some basic physical tasks and talk to people. We're interpreters and facilitators, but the knowledge-based judgements and decision-making is best left to the higher intelligence. The short-term problem is that all of the knowledge-work that humans currently do will be redundant, and the only thing remaining that humans can do better than an AI is, as he puts it, "[the] physical implementation of many AGI decisions." And he's worried about the economy! Does he not realize that he is going to need to find a new job delivering packages because the AI could have written this?
I feel like the expectation of GDP growth is one of those things that might be using the wrong language for the thing being discussed. For example, today I have on my phone what Edward Bellamy called in 1887, "the limit of human felicity" referring in this case to music on the demand of any kind I want for a subscription fee of about 41¢ a day, along with video on a wide variety of interesting subjects, without commercials. In the GDP graph this change means nobody buys CDs any more, but in the experience of personal abundance this renders
A tutor program that functioned on level with an actual human tutor would probably cost less per month than tutoring costs per hour. AI that masters making 3D models that are fully functional or that can animate artwork could transition us from the "major projects of billion dollar studios" to the "modern publishing quantities" for movies. I think you're right that we're unlikely to solve the substantive problems of the world with an early AGI but, even with just good AI, you might genuinely increase artist productivity by orders of magnitude while adding basically nothing to GDP. I'm not familiar enough with other intellectual work fields to say if it feels like there are things around the corner for them but it could easily be similar. That stuff might not show up in GDP but it would definitely show up in people's experiences of the world and we would all be much "richer."
None of that feels remotely dystopian. I'm honestly hyped. But it does feel like de-skilling might soon blow through the intellectual halls just as it did the artisanal classes in the 1870s-1900 over the next thirty years. I don't think we'll enjoy that process very much, especially in heavily skilled countries like the United States. If the history of the conduct of the artisanal classes is any hint, we'll be in for (or perhaps already are on) a wild ride.
AI is both amazing and a threat in so many ways. How long before (if already) will a Terminator evolve. We have fighter aircraft now equipped with remote piloting. It is a very small step for a totally self-directed and dangerous if not sentient machine! Joseph is mostly optimistic but I am not quite as comfortable given the nature of our species!
What good would increased productivity bring to the average worker if most of the benefits flow to the top owners and Wall Street bankers? Since the 70's this has been happening and salaries have not keep up with housing, healthcare and other basic needs. What has increased is the reliance on borrowing.
I don't know why economists can't figure out that there is a limit to humans abilities to consume. There are only 24 hours in a day and you have to sleep for 8 of them. Consumption can be maxed out and in fact you see it with wealthy people all the time. They end up pouring their money into assets or charity.
Great Article from various perspectives & context. My #1 hope of technology period and especially AI is that over time (Long-Term) it helps "Root-Out_Tribalism" in American Culture. In today's culture "Trust" is the issue. I say to the trust issue is something that should never change in any endeavor..Individual-Situational-Awareness...use it. The process of validating the veracity of information “NEVER” stops! Gonna be a great "Tool" for productivity and the advancement of society!
Truck drivers are safe. Absent dedicated "AI driving" roads/lanes, absent a multiple-order of magnitude increase in accuracy, it's not happening anytime soon. It'd be cheaper to improve the railway system.
Artists are less safe, but only the marginal ones. Many artists have already adopted AI as a tool.
Lawyers, yeah. These guys are f*cked. Most of what they do is parasitic anyway, and at best a necessary defensive measure almost indistinguishable from insurance.
As a student, ChatGPT specifically has caused me a lot of mental turmoil since I became aware of it before the end of last year. My perspective on my own capabilities shifted for the worse and I’ve consistently doubted my ambitions. Thank you for reminding me of my innate human value - better yet through the lens of an economist! More young people need to hear what you have to say!
At some point, AIs will speed-up R&D to superhuman levels of speed, creating new instances of AIs will be way cheaper that forming people to do those jobs and each will perform many tasks at superhuman-levels of speed (and you can increase those features by adding more GPUs), at that point, it might just be that all of the 'physical' jobs will end up being automated without too much delay.
Great article! Particularly about the future roles of human labor due to our comparative advantages with respect to AI.
I believe another important job for humans to perform is providing responsibility for decisions, including culpability for bad decisions. Eg, AI can already answer many common medical questions with comparable accuracy to human doctors. [1] Yet patients will demand human responsibility for major medical interventions, and even for the decision to do nothing. That ensures that there is someone to fire, sue or jail for egregiously bad decisions.
In looking at the future of work, I think we can take inspiration from the role of Vice President at financial firms; a relatively junior role that can make up the bulk of the employment for many key divisions. It’s my understanding that this is not just title inflation for the sake of workers and customers but instead a legal necessity. Many actions taken on behalf of a bank can only be performed by an executive.
Even when a firm itself can be culpable, and thereby sued and fined, we humans simply demand a specific person take responsibility for important decisions. With AI bringing massive amounts of information and intelligence to task, our society can now make many more decisions with higher quality. Yet we'll still want a human to blame when things go wrong.
[1] https://arxiv.org/pdf/2212.13138.pdf | summarized in https://twitter.com/emollick/status/1610261628607512576
Love this - very interesting
This is by far the most dangerous invention of mankind, there is no way this will work out well for humanity. Look at what has happened to our society already due to unrestricted and unregulated growth and utilization of technology. Once again this will funnel all the money and power to a few technocrats and all of humanity will be their slaves, just like now only worse. The internet should have been subjugated to common decency standards like the rest of the worlds media and chat ai should never be made publicly available. Is it reasonable when each of us has access to something that can easily wipe humanity of the face of the earth, I don’t think so, it’s like giving everyone the nuke launch codes and waiting for me to launch them, if they have a bad day it could happen. The only way to allow this technology safely is to legislate strong legal and tax obligations on all robotics and technology to protect society from what has already happening. For every machine that displaces a human the owner of that machine should be required to pay the equivalent tax to cover the cost for supporting the human. Social security, unemployment, medical care and retraining if possible should all be the the responsibility of who ever disrupted the system by elimination of jobs. Humans must have jobs to function normally, only a very few can stand to not do something. This should be the first priority, keep humanity engaged in life is the crux of life. If people are unable to see a way to the future they will give up and that’s what’s happening now. It’s hard to see how this technology will benefit mankind for more than just a handful of the wealthiest citizens who don’t need any more of anything least of which is control over the rest of us.
I don't see how humans could ever have a comparative advantage with AI in a world where AI can do every task better than humans. You can use ChatGPT to write articles for more than 99.9% less than a human writer. If ChatGPT was better than human writers, there wouldn't be any humans willing to work for 99.9% less than minimum wage to try to compete on price. In any industry where AI has an absolute advantage & it's legal to replace humans with AI, human labor is worth nothing. It's certainly possible that we AI never gains an absolute advantage in certain industries. Childcare and elderly care seem like industries where people may always prefer human labor even if they cost hundreds of times more than robots, but that means humans have an absolute advantage in those industries.
So this guy's bias is basically 'concern for the economy and investing', he's really just asking the question "will global GDP fall as a result of AI-driven unemployment", and his thesis is that "no it wont because humans will continue to want more things and have higher expectations" and so the "services economy will rise". Like, assuming human services will still be considered important in the world where an AGI is making all the decisions, and human consumption will continue to be affordable (it wont be if mass numbers of people are quickly redundant). Also kinda missing the point that physical work is the first thing we've tried to automate because no one wants to do it. Great, so our software engineering, financial analyst, doctor, and lawyer jobs will go, but there will be plenty of jobs for us in delivering packages to peoples front doors, and having conversations with other humans about the surgery the AI recommends to patients. So we dont need knowledge anymore, we just need to be able to do some basic physical tasks and talk to people. We're interpreters and facilitators, but the knowledge-based judgements and decision-making is best left to the higher intelligence. The short-term problem is that all of the knowledge-work that humans currently do will be redundant, and the only thing remaining that humans can do better than an AI is, as he puts it, "[the] physical implementation of many AGI decisions." And he's worried about the economy! Does he not realize that he is going to need to find a new job delivering packages because the AI could have written this?
I feel like the expectation of GDP growth is one of those things that might be using the wrong language for the thing being discussed. For example, today I have on my phone what Edward Bellamy called in 1887, "the limit of human felicity" referring in this case to music on the demand of any kind I want for a subscription fee of about 41¢ a day, along with video on a wide variety of interesting subjects, without commercials. In the GDP graph this change means nobody buys CDs any more, but in the experience of personal abundance this renders
A tutor program that functioned on level with an actual human tutor would probably cost less per month than tutoring costs per hour. AI that masters making 3D models that are fully functional or that can animate artwork could transition us from the "major projects of billion dollar studios" to the "modern publishing quantities" for movies. I think you're right that we're unlikely to solve the substantive problems of the world with an early AGI but, even with just good AI, you might genuinely increase artist productivity by orders of magnitude while adding basically nothing to GDP. I'm not familiar enough with other intellectual work fields to say if it feels like there are things around the corner for them but it could easily be similar. That stuff might not show up in GDP but it would definitely show up in people's experiences of the world and we would all be much "richer."
None of that feels remotely dystopian. I'm honestly hyped. But it does feel like de-skilling might soon blow through the intellectual halls just as it did the artisanal classes in the 1870s-1900 over the next thirty years. I don't think we'll enjoy that process very much, especially in heavily skilled countries like the United States. If the history of the conduct of the artisanal classes is any hint, we'll be in for (or perhaps already are on) a wild ride.
AI is both amazing and a threat in so many ways. How long before (if already) will a Terminator evolve. We have fighter aircraft now equipped with remote piloting. It is a very small step for a totally self-directed and dangerous if not sentient machine! Joseph is mostly optimistic but I am not quite as comfortable given the nature of our species!
What good would increased productivity bring to the average worker if most of the benefits flow to the top owners and Wall Street bankers? Since the 70's this has been happening and salaries have not keep up with housing, healthcare and other basic needs. What has increased is the reliance on borrowing.
I don't know why economists can't figure out that there is a limit to humans abilities to consume. There are only 24 hours in a day and you have to sleep for 8 of them. Consumption can be maxed out and in fact you see it with wealthy people all the time. They end up pouring their money into assets or charity.
"no laws will be passed to ensure that, for example, your driver, lawyer, cook, doctor, or financial advisor must be a human being."
You don't think lawyers in Congress and legislatures everywhere won't act to protect their turf?
Great Article from various perspectives & context. My #1 hope of technology period and especially AI is that over time (Long-Term) it helps "Root-Out_Tribalism" in American Culture. In today's culture "Trust" is the issue. I say to the trust issue is something that should never change in any endeavor..Individual-Situational-Awareness...use it. The process of validating the veracity of information “NEVER” stops! Gonna be a great "Tool" for productivity and the advancement of society!
Truck drivers are safe. Absent dedicated "AI driving" roads/lanes, absent a multiple-order of magnitude increase in accuracy, it's not happening anytime soon. It'd be cheaper to improve the railway system.
Artists are less safe, but only the marginal ones. Many artists have already adopted AI as a tool.
Lawyers, yeah. These guys are f*cked. Most of what they do is parasitic anyway, and at best a necessary defensive measure almost indistinguishable from insurance.
As a student, ChatGPT specifically has caused me a lot of mental turmoil since I became aware of it before the end of last year. My perspective on my own capabilities shifted for the worse and I’ve consistently doubted my ambitions. Thank you for reminding me of my innate human value - better yet through the lens of an economist! More young people need to hear what you have to say!
At some point, AIs will speed-up R&D to superhuman levels of speed, creating new instances of AIs will be way cheaper that forming people to do those jobs and each will perform many tasks at superhuman-levels of speed (and you can increase those features by adding more GPUs), at that point, it might just be that all of the 'physical' jobs will end up being automated without too much delay.
https://towardsdatascience.com/the-carbon-footprint-of-chatgpt-66932314627d
"...ChatGPT is based on a version of GPT-3. It has been estimated that training GPT-3 consumed 1,287 MWh which emitted 552 tons CO2e."
https://les-enovateurs.com/chatgpt-4-faster-more-powerful-smarter-and-even-less-eco-friendly
We need more data centers and more power 🤷