7 Millions Jobs Vanished Overnight...

So welcome back to post AGI Economics. In today's video, we'll be taking a look at some of the most important stories regarding economics in AGI and the evolving job market with regards to how artificial intelligence will impact the evolving landscape and the future of the economy. Now, there was a recent podcast that was launched by Alexander Madri, the head of preparedness at OpenAI. In this episode, he dives into a lot of key topics with a leading economist and Eric is a professor at Stanford University. This is There are two key things from this interview that I watched that I wanted to share with you all because I think it shows us where we are headed in terms of the AI development landscape affecting the economy. The first thing that they actually do talk about is AI being a complementer versus it being something that takes all jobs.

There's a very important concept there about AI being not the substitute, but complement. And you talk about things like touring traps or could you tell us more about.

Yeah, thank you for highlighting this, because I think probably the biggest misconception that I run to as an economist is this assumption that AI is going to take all the jobs and that AI is a perfect substitute for humans. And it's true that sometimes AI does substitute for people, but more often than not it ends up being a compliment. Substitute is when something replaces something. Complement is when something makes it more valuable. If you look at through most of history, technologies have been both. But the more important thing has been AI as a complement. And one way to see that is that think of how valuable is one hour of human labor? It's gotten more and more valuable over time. In 1700, it really wasn't that valuable. In 1800, 1900. Today it's more valuable than ever before. Why? Because it is complemented with technologies. If you're digging a ditch, you have a bulldozer or steam engine that can lift things. If you're writing software, you have computers that can help you create software faster than ever. So an hour of labor can create far more value than before because of these technological complements. It's also true for AI that in most businesses, like in the call center example, you get more value by using the AI to complement human workers than trying to replace it. And part of that is because, brilliant as the machine learning folks are, they haven't been able to make machines that can do the full set of things that humans can do. And so typically when you're trying to solve a problem, AI can do parts of it, but humans are still needed for other parts of it. Like in the call center example, we looked at all the different questions that were asked, and there's some questions that are asked very frequently. How do I change my password? I'm locked out. And there are other questions. There's this long tail of questions that only appear once in the data. Some complicated tax question or something that just has never been seen before. It turns out not surprising that machine learning is good at things where there's a lot of data. You need data to do machine learning. So those common questions, they have some really good, crisp answers for, but that long tail of unusual questions, AI machine learning can't help much with those.

So in this clip, this is basically.

Where he's talking about how when you have an AI system that is quite good at, you know, processing lots of data, this does work. But because AI is currently generative, its architecture isn't neuro symbolic. It's not able to, you know, solve issues that, that are pretty much unseen. On the tail end of the distribution tasks, for example, you know how. And I'll give you guys this example, which is pretty much tying into his example. If you've ever called up a line and you know how it says, press 1 if you want to do this, press 2 if you wanted to do this. I guess you could kind of say that LLMs are kind of an extended period of that, because when you do get into certain specific issues, LLMs truly cannot help with that because the issues are far too specific. And, and it's a novel case. And of course, in the future, I do think certain technologies are going to be developed on top of LLMs. There's going to be new architectures that are going to, of course, solve this issue, maybe even synthetic data to where they're going to have the AI maybe handle unseen cases and then solve those unseen cases. So it's going to be trained on like 10,000 or 100,000 unseen cases. Maybe that's how the future could possibly be. But I do think that what he's stating here is a rather important factor. A lot of the times most people do think that it is going to replace all jobs. But I would also argue that in this podcast, one thing that they did also cover was the fact that it is very, very hard to predict because of the uncertainty of the future. And another thing that I do want to show you guys is them talking about the future in terms of what to study. One of the key things that many people have been confused about is which careers are going to be there and which careers aren't going to Be there in the future. Making predictions like this is really hard. As I said in the previous episode of Post AGI Economics, even Geoffrey Hinton was wrong about his prediction of radiologists. Now towards the end of the podcast, they actually do talk about this future.

Of like, what are the right jobs to be in during this AI transformation? So my, both of my daughters, they soon will enter high school and I'm really at loss. What should I advise them? Like what career to pursue them. So what would be your advice here?

Well, my advice is be careful about any advice from me because it's just tricky. I'm becoming humble. In the second machine age. I had this vision like a lot of people did, but I kind of Andy McAfee and I wrote about it, where technology was good at sort of routine, repetitive tasks and more creative work that was something that humans were better at and more interpersonal work was something that humans were better at. And I think that was kind of true then and was true for a number of years. But now LLMs and generative AI is flipping that around and a lot of creative work can now be done by machines or somewhat creative and increasingly creative. So my new advice is be flexible. I mean, is that be ready to make changes and don't get locked in. Don't think that, okay, I'm going to be a coder, or I'm not going to be a coder or I'm going to be an English major that be adaptable. And I think that that may be where humans have an advantage. I was having a conversation with Reid Hoffman the other day and we were talking about this. I mentioned this long tail of tasks.

And I do want to say his advice here is actually the same advice I've given in my community and one of the frameworks that I've actually developed and worked on for quite some time. One of the things that we talk about is being adaptable, agile. And this is really important because as I've said before, the future landscape is uncertain and it's dynamic, meaning that the future area where you might think that this area or this career is going to be 100% certain and it's never going to be automated away.

You know, you have to be open.

At least somewhat to the possibility of being dynamic. So by that, you know, you should be focusing on what strengths you have, what weaknesses you have, what areas you're comfortable in. Of course, what industries are going to be expanding that have complementary skills. These are kinds of some of the things that I've been looking at, some of the things that I've been talking about with my community. And it's refreshing to hear that this perspective is reiterated by someone who holds a major position when it comes to evolving the discourse on how economics and AI is going to evolve. And whilst yes, the perspective of changing entire careers might seem daunting at first, I think the nature of humans is to be very adaptable. And I do think most humans can adapt well if given enough time. There was also this 7 million rideshare drivers jobless overnight as China's foolish autonomous cars snatch jobs. I do want to state that this title is quite exaggerated in the sense that it didn't immediately put rideshare drivers jobless overnight. And of course here the comment here says I can confidently say that overnight is dumb and inaccurate. The rollout is still very gradual to certain areas of the city at times. But I think what this shows us is that the impact of this is going to happen. And the reason I'm making this series on my channel, the reason I'm talking about these things, is because I want people to understand that it doesn't make sense to just stick your head in the sand and be like, okay, AI is coming for my job, but I'm just going to pretend it's not so that I can feel comfortable about the situation. It makes sense to actually look at what industries are changing and evolving so that you can actually put yourself in a better position. Because I can guarantee you guys there are going to be a lot of great opportunities out there that are going to be needing individuals like yourself. And I'm sure that if you're someone who's in the area of AI now, especially right now, since it is currently summer, as I'm releasing this episode, the AI hype tends to die down quite a bit. And, and if you're still watching these videos, it means that you have a heightened interest in AI, which means that you're likely to be someone who finds out about things first, which means you can take these opportunities first. So essentially this piece of news is where China's robotaxi push sparks concerns about job security for drivers. And this is one thing I did talk about before, about how if we take a look at what Tesla's long term plans are, and if we take a look at what some of companies like Waymo are doing, these companies are effectively entering the same job market as Uber drivers. Now, I'm not sure that many people think of Uber drivers as a career, but I do think it is a way for certain people to earn an extra wage outside of their normal working hours. Now the article dives into the fact that China's year long effort to develop robo taxis is starting to gain traction with consumers, but it is also rattling taxi drivers worried about losing their jobs as a result of increasing competition. Baidu's robo taxi unit Apollo Go became one of the top 10 trending hashtags on social media platform Weibo yesterday amid reports of rapid user adoption in Wuhan City where the company began operating fully staffless vehicle in certain districts in March. China had more than 7 million registered ride hailing drivers as of the end of May, roughly twice as many versus the 3.51 million drivers as reported for July 2021. Now I understand the concern here with the economy in a strange point to where inflation is continuing and we do have a situation to where people are trying to earn more money as the cost of living increases. Jobs that provide some security and extra money are now under some kind of threat. And I do think that if you pay attention to around the world this is going to happen. I recently saw articles to where there are certain legislations being paved and certain laws being passed to allow robo taxis in the UK and of course in America. And what's interesting here is that that in may more than 70% of Apollo go robo taxi rides in April were fully driverless with no human staff. In September side he predicted that that share would reach 100% in the coming quarters and allow Apollo Go to break even in Wuhan first. One thing that I've always wondered is is this stuff going to be widely positively received in America? I think for the large majority it will be. I've seen a lot of tweets, a lot of comments and a lot of reviews of the Waymo. If you don't know what this is, it's basically a robo taxi, an autonomous self driving car driven by AI that just takes you to your destination and it's currently being rolled out across the United States over the next few years. Now I do wonder if people will take likeness to this. Maybe some states will ban it. I wonder if New York is going to consider the fact that it has its taxi drivers there. Are they going to allow waymos in New York? These are all the kinds of things that depending on where you live you might need to think about if it does affect you. And if you're wondering why I'm asking that question. Well some people actually set a Waymo on fire. So I do think there is always going to be a little bit of resistance with these new technologies. Next we have an article from Forbes talking about can universal basic income save us from the destabilization of AI and automation? As AI and automation continue to reshape the global job landscape, societies worldwide are grappling with profound economic and social changes. Major recent job layoffs, such as those announced by companies like Tesla, Google, Microsoft, Nike and Amazon, highlight the urgent need for solutions. Nearly 40% of business leaders anticipate layoffs this year, with around 4 in 10 planning to replace workers with AI. And I do think that once future systems are released and certain demos of certain capabilities in terms of reliability, speed and performance are announced, I think this is going to truly show you all how quickly things are going to change. Of course, here there are questions. Will there be enough jobs to go around? And who or what will hold those scarce positions? The ability of AI to perform current human job functions is leading to job displacement. It also talks about here how Guy Standing, a renowned economist and the author of the Politics of Gaining Control in the Age of Uncertainty, emphasizes the destabilizing effects of economic inequality and insecurity. The AI revolution is accentuating the flow of income and power to the owners of the property, leaving a new class, the Precariat, wallowing in insecurity and existential fear. He argues that UBI can provide a financial safety net that mitigates these issues, allowing individuals to pursue meaningful activities and reduce mental health problems stemming from economic stress. One thing that most people don't realize is that 90% or 70% of their actual stress does come from not having enough money to make ends meet. This isn't just a stat that I pulled out of a random book. This was based on a study that actually evaluated the psychological performance of those who had a regular steady income through property. It showed that their stress levels were completely down and they tended to live longer. You can see here that they also point to the social and health benefits demonstrated by the 1970s Mincome experiment in Canada, where participants were 8.5 less likely to be hospitalized and a basic income could reduce healthcare costs significantly. That is also something I didn't think about. If there was a lot better mental health and a lot better conditions for people to live in, would it put a lot less health care systems in less of a strain? And I think it actually makes sense now. Do I think UBI is going to come out very soon. I mean, I do think that there probably will be small pilot programs continually tested across America. And I do think that as these systems get more and more capable and as companies start to adopt them, I think the world might struggle to find places in which humans are increasingly valuable. And and I do also think that whilst yes, humans are adaptable, not everyone is used to adapting as quickly as possible, and some people might struggle to transition from certain careers if they haven't prepared in advance. It's one thing to be in my community or watching these videos, understanding the landscape, being ahead of the competition and preparing yourself, but many people just aren't paying attention to AI, so are in for a wild surprise when certain future capable systems are released and their entire industries are shifted and impacted. This is why I'm advocating and telling you all now that it's better to think about solutions in the early days whilst you still have a lot more time. There was also this article recently released where it says Silicon Valley Libertarians ponder Universal basic Income as AI wiped out jobs Dramatic job and salary destruction will accompany the spread of AI, prompting big tech leaders to support trials of guaranteed income models to lift displaced workers from economic despair. The initial promise of early results from AI taking over some workplaces suggests that the tech will will likely to be among the most transformative innovations since the Industrial Revolution. And much like that historic upheaval, there is a non zero chance that will entail dramatic job destruction as computers take over work from human workers, allowing companies to vastly multiply efficiencies and profit. Now with the most conceptions about UBI being something to finance cigarettes, alcohol and tattoos, which are many of the critics talking points. According to Guaranteed Income Pilots Dashboard that compiles data from 30 United States states UBI test programs, over 43% of monthly payments are typically spent on essentials including food, housing and healthcare. Transport costs accounted for 9.4% of UBI recipients. Spending and discretionary retail services were the largest category at 36%, meaning a lot of de facto essentials may have been included under that heading too. So basically they're stating here that this is not true. The majority of people who are poor and don't have enough money that actually need UBI won't actually be spending it on alcohol, cigarettes and tattoos, which is pretty obvious. Now. One thing that I found interesting about this article is that it does include some statements from Sam Altman. It says OpenAI CEO Sam Altman, whose text stands to be hailed and blamed for the labor transformations it generates. He's aware the heightened productivity AI promises will also destroy many jobs and salaries, which may require governments and companies to provide at least minimal income to fill the gaps created. It's impossible to truly have equality of opportunity without some version of guaranteed income. This is what Sam Altman said way back in 2016, according to the Times. The paper also noted the OpenAI boss ponied up $14 million of his own cash, a huge $60 million data driven study on UBI programs and that's out next week. So if you remember, Sam Altman previously spoke about how they were doing a very large UBI study because they wanted to know the impacts of how this technology is going to impact the economy when it displaces many jobs. And apparently that is out next week and I can't wait to see the results of that study. You can see here Elon Musk says probably none of us will have a job, musk said at a test conference in Paris where his views on guaranteed revenues were retrofitted to his rightward political evolution, which we won't have universal income, we'll have universal high income. Now, this article basically touches on some of the points that we've spoken about previously, but I think the main thing here is we all know the direction we're heading in. And like I've said before, one of the main concerning things that I've seen as a trend, as someone who's looking into this, that there isn't that many answers when you do truly look at some of the scenarios because we are entering uncharted waters. By that I mean the very best thing you can do is be as adaptable as possible so that if something does occur you are ready for the change. It's better to be prepared and not need to be prepared than to not be prepared at all. Now, a point that I do want.

To close on here. Whilst yes, this video title might have actually been somewhat clickbait, I think video titles like this wouldn't be out of.

The realm of reality if these companies weren't actually paying attention to the economic landscape. One of the things that OpenAI have.

Actually done to prevent titles like 7 million people jobless overnight is their stance on iterative deployment. OpenAI knows that the technology they are.

Currently developing, with every iteration it manages to encapsulate and cover a wider range of different tasks. And with this they know that there.

Is going to be a lot of economic displacement. I think it's truly a good thing that OpenAI are doing by releasing this technology in stages. But what that doesn't mean is that it doesn't mean that all of those industries are not being affected. For example, just because this robo taxi thing is being rolled out slowly. It doesn't mean that in the future there isn't going to be a hyper.

Competitive environment where a lot more people.

Take autonomous vehicles instead of those with.

Actual drivers in them. The point that I'm trying to make here is that whilst yes, people aren't going to be jobless overnight, I still.

Think there is going to be a slow increase in the amount of people that are going to be unemployed in certain industries and it's going to be a slow churn, one that you might not realize. You're basically going to be a frog.

In a boiling pot. Which is why I'm making this video.

This piece of information and this is what I talk about all the time. And stuff like this is where people don't look at the nuances of how different things affect different industries. So a $3 billion software unicorn scrapped.

Its plan to give AI workers rights.

After tech executives said it disrespects the.

Humanity of your real employees.

Human resource software company Lattice learned the hard way that employees are not quite ready to work alongside AI counterparts. The whole point of this article, which I'm going to tell you guys about, is the fact that humans fundamentally in some industries decide whether or not AI gets used, because humans are somewhat of the not in this case, end users of the final end product. So it says. The tech unicorn founded by Sam Altman's brother Jack Altman, valued at $3 billion in 2022, announced it made history this week by giving so called digital workers employee records and integrating them into organization charts, allowing its human co workers to see the roles of AI in the workplace. And after a surge of online pushback, the company said it would no longer pursue the project. Only three days after the initial announcement. You can see here, basically these guys launched AI agents in their workplace that were actually doing work. The humans could actively see them. And humans just literally said no, we don't want this. You can see here that the article continues to state that this strategy and messaging misses the mark in a big way. And I say that as someone building an AI company, treating AI agents as employees disrespects the humanity of your real employees. Worse, it implies that you view simply humans as as resources to be optimized and measured against machines. And there's already mounting anxiety around the role of AI in the workplace, particularly as its use proliferates across different industries. Of course, in March 2023, Goldman Sachs reported and found that AI could replace or degrade up to 300 million jobs in the United States and Europe. Whilst venture capitalist Kai Fu Lee predicted the 50% of human workers will be replaced by AI by 2027. And of course, Lattice's own transition to AI may have not relaxed those anxieties. So whilst yes, AI advances, I think these companies are going to have to think really carefully about how they integrate.

AI into the workforce if they're not careful. As I've said time and time again.

This is going to have more of a negative impact on your company rather than a positive one. And this is something that I'm doing myself by being very careful and very selective and to where I use AI, because I do realize that whilst yes, it is efficient and effective, it's not good for absolutely everything. If you enjoyed this video, hopefully you.

Have a wonderful day.

If you have any questions, leave them.

In the comment section below.