FOR EDUCATIONAL AND KNOWLEDGE SHARING PURPOSES ONLY. NOT-FOR-PROFIT. SEE COPYRIGHT DISCLAIMER.

AGI Will Freeze the Economy | Matthew Berman

This video is brought to you by Augment Code. Will reaching AGI lock you into a permanent underclass. More and more I’m hearing that from people in Silicon Valley, people who work with artificial intelligence and I had to dig into it. I had to better understand it because if this is true, it has major societal implications. Let me break this all down for you. So first AGI artificial general intelligence is when AI reaches the point that it is better than humans at knowledge work. So developers, customer service, data entry, data analysis, basically anybody who’s sitting in front of a computer. And when this happens, the only limitation to knowledge work becomes how much compute you have. And what does that take? Capital, money. You have to buy compute. And all of a sudden the major piece of leverage that humans have had for history which is our labor suddenly becomes valueless. And at that point the only thing that matters is how much compute you can buy to run your AGI. So wherever you are socioeconomically at that moment that’s where you’re going to stay. At least that’s what more and more people at the frontier of AI seem to be thinking. And so if you look at the major AI companies, they are all investing so heavily into building out these massive server farms because they understand at the end of the day, compute is going to be the currency of the future. And so they’re all rushing to build it out because they don’t want to get left behind. And once AGI is achieved, capital owners will essentially be able to replicate talent infinitely. AI talent doesn’t care who its boss is, doesn’t care where it’s working, doesn’t care about perks. The only thing AI workers care about is electricity going through GPUs. And whoever can pay for that will be able to hire these AI workers of the future. But what does that world look like when human labor loses its value in the economy? With no incentive for companies to invest in human development, societies risk sliding into stasis. Wherever you are is where you stay. This is the big fear of the permanent underclass. And so this concept of the permanent underclass isn’t necessarily about poverty. It’s about lack of mobility. If you’ve watched this channel, you know I tend to be very optimistic. But in this video, I really want to show you the other side of the argument. And at the end, I am going to give my optimistic take. But you’re already starting to see some of this. Now, entry- level jobs are becoming much more difficult to land as a recent college graduate. And so, there’s this culture of urgency. And I’ve heard this from numerous people meeting them in person saying, “I have to accumulate as much capital as I can over the next 12 months, 2, 3, 5 years before AGI happens.” Because once AGI happens, whatever I have is what I’m always going to have. And these aren’t just fears with no data to back them up. Let me show you a few things. First is Stanford’s paper called Canaries in the Coal Mine, which is a study of ADP payroll data through mid 2025, and it finds about 13% relative employment declines for 22 to 25 year olds in the most AI exposed occupations. That means if your job can be done by AI, recent grads that are trying to find entry-level jobs are finding it more difficult than ever. But here’s the thing. Older workers in the same roles are flat or growing, and wages have not broadly fallen. So, there’s some decent news there. Next, career networking website Handshake put out a study showing that labor demand for entry levels is down. 15% fewer entry-level postings and 30% more applications per job yearover-year for the class of 2025. Okay, so all of that looks bad, but also all of this infrastructure investment is going to mean tokens and inference and compute gets a lot cheaper, right? Well, yeah, it definitely will. The cost of running AI is essentially going to approach the cost of the chips plus electricity to run through them. And once that happens, AI is going to be incredibly inexpensive. So that’s good, right? Now everybody gets AI and there’s no more permanent underclass. Well, pause for a second. There’s this thing called Javon’s paradox. And if you’ve watched this channel, you’ve definitely heard it before, but it basically means as some resource gets more efficient and less expensive, we actually tend to use more of it overall. So although the cost per unit of some resource, let’s say electricity or compute, decreases, the amount of use cases that are now economically viable increase. And so the total spend increases. So we’re kind of back in the same spot, the potential for a permanent underclass. But some of the smartest people in the world are telling us three letters are going to save everything. UBI, universal basic income, and in some instances, universal high income. People like Sam Alman and Elon Musk seem to believe in UBI and UHI. So, if you’re not familiar with universal basic income, it is a social and economic policy in which the government gives its citizens regular unconditional cash payments to pay for the basics in life. Rent, food, health care, things like that. Now, of course, Elon Musk believes in universal high income. Not just paying for the basics, but essentially having enough money to choose what you do with your time. But, of course, there’s a counterargument for UBI. A lot of people think it’s just socialism and it may cause inflation. We’ve spoken with the expert Scott Santins and he actually has shown that UBI is actually a great motivator. When you have more free time on your hands, you actually tend to be more entrepreneurial. That means you’re going to go out, you’re going to look for problems, you’re going to start a business, potentially hire more people, and these are all really good things. But that doesn’t quite answer the question of why isn’t AGI still just going to do those jobs? I again am optimistic. My optimism comes from something Biology said pretty recently. AI is likely going to be only capable of completing tasks middle to middle, not end to end for the foreseeable future. I do believe for a wide range of tasks, humans are going to be required in the loop. a human is going to be required to prompt AI and then maybe in the future orchestrate swarms of agents and then of course when the output comes we need to review and verify it and so although the middle is going to be pushed outwards and AI and AGI is going to be able to accomplish more and more tasks there is still a place for humans on the ends of these tasks and the total number of tasks that can be completed and require a human on the ends is going to increase dramatically and So what we’re going to see is pretty fundamental shifts in the actual work that humans are doing, but humans are still going to be required. This is called job augmentation, not automation. And speaking of augmenting, the sponsor of today’s video, Augment Code, is a perfect example of something you should use and learn right now. You’re a professional software engineer, and sometimes Vibe Coding just won’t cut it. Augment Code is an AI assistant built for real engineering teams. It ingests your entire codebase. I’m talking millions of lines, mature code bases, tens of thousands of files. So every suggestion that it gives you lands in context and keeps you in flow. And with Augment’s new remote agent, you can ceue up parallel tasks like bug fixes, features, and refactors. Close your laptop, go do something else, come back, and it’s done. And the only thing you need to do, review the poll request. So, if you are reaching the limit of your other AI coding tools, give Augment Code a try. Augment Code also never sells your code, so your team’s intellectual property stays yours. And you don’t even have to learn new tooling or switch to something else. You can use Augment Code within VS Code, Jet Brains, Android Studio, or even Vim. Start your 14-day free trial at augmentcode.com. Let him know I sent you. All the links down in the description below. Thanks again to Augment Code. Now, back to the video. I interviewed Aaron Levy, CEO of Box, and he said the same thing. Check this out. The mistake that some companies make is to say, you know, go use AI, and where you can’t use AI, that’s where we’re going to add headcount to the organization. I I generally think you want to kind of do the opposite where you basically say the parts of the organization that are using AI will be the ones that get more budget. If you can actually make your organization more efficient, you’ll probably be higher on the list of the teams that get next set of headcount or the next set of budget to go deploy against problems. And here’s another follow-up clip in which I asked him a few months later the same question. Does he still believe it? Let’s watch. This industry moves very fast, but fortunately no change in the past like two months on that answer. Um mo mostly uh I think AI is largely going to be deployed at areas where companies have previously been constrained by resource constrained by their ability to execute which will in many cases actually lead to faster growth of companies which ironically could even mean more hiring because now you need people to go and handle the next set of bottlenecks that you’re bottleneck on. So if you think about like what like a hyper growth uh software startup maybe they can code more efficiently but all of a sudden they have more customers they have to go serve which means more people in sales or consulting or customer success. So I actually think this will lead to growth in just different areas. And here’s the CTO of Open AI essentially saying the same thing. Take a look. I think the edges will become thinner. I mean we still have hallucinations right? It’s not perfect in everything and it can create mistakes etc. So I think a lot of times the answer is going to depend on what is the error bar that you have for the task you’re trying to do and can you accommodate that. There is also going to be tremendous job creation jobs that we can’t even think of today. Think about my job. Speaking in front of a camera to people on the internet didn’t exist 20 years ago, let alone a 100red years ago. People couldn’t even have imagined it. And there’s going to be a lot more jobs that today we can’t even imagine that are going to be created because of AI. There is also this tremendous competitive opportunity for new founders to go take on these large companies to be able to do so much more with so much less because of artificial intelligence. And AI is going to allow more people, individuals, to be able to do so much more creatively as well. Here’s an interview with Reed Hoffman I did with vibe coding and everything else. There’s going to be a lot more like like essentially software for just this task. Right. Right. That’s part of where the infinite demand comes from. For example, when you think about create creativity like creating software for CG for movies, well, that was super expensive and so only a small number of very high-end, very big budget Hollywood films can do that. Now, the like for example, documentaries are going to have access to all this because of vibe coding. All right, so let’s wrap this up. First, there is going to be massive transitions in the jobs that we do today, knowledge work specifically, and there’s really no avoiding that. We are going to have to be proactive to teach people how to use these tools, AI tools, to be most productive and to be most valuable in the workforce. If you’re watching this channel, you’re already a step ahead. Plus, it takes time to implement this stuff. Just because the models are capable doesn’t mean we actually have the infrastructure in enterprise to actually go execute the tasks that we need to. AGI and artificial super intelligence are coming. But when they come, it’s also going to take time to really roll it out in society. Next, I do believe in augmentation. I think some jobs are going to be automated completely, but a lot of jobs are simply going to be transformed into looking just different. There is a chance that universal basic income works really well. There is a chance that universal high income is not only possible but obviously works really well also and these things are because of all of the abundance we have because of our productivity due to AI. So is AI going to be like other major technological shifts in history or is this going to be different? Let me know what you think in the comments. I believe in a future in which AI is going to help everybody be more creative, be more entrepreneurial, accomplish more. We’re going to solve disease. We’re going to have mass abundance. It’s going to be amazing.

FOR EDUCATIONAL AND KNOWLEDGE SHARING PURPOSES ONLY. NOT-FOR-PROFIT. SEE COPYRIGHT DISCLAIMER.