The year is 2030. OpenAI has just released its o7 model which they claim is smarter than any human on Earth. Artificial General Intelligence (AGI) has been achieved and we have just entered the world of Artificial Super Intelligence (ASI). There is no intellectual task that can be done by a human that can’t be done even better by an advanced LLM. A market rally ensues.
***
If I had shown you this passage a few decades ago, it would have read like something out of a sci-fi novel. But today, there are many who believe that such a future is on the near horizon. Of course, I am not qualified to discuss whether AGI/ASI will ever come to pass (or when). But, let’s assume they are right and that we are only 5 years away from such an event.
In this scenario, what will it mean for humans? Some believe that the creation of AGI will reduce the need for humans and human labor, making many of us unemployable. If this ends up being true and capital ends up crowding out labor altogether, how can we prepare ourselves for this coming economic upheaval? How do we invest for a post-AGI world?
In this post, I will propose three possible solutions to this problem:
- Just Own the Damn Robots (Part 2)
- Hedge Your Labor, Not Your Portfolio
- Stay the Course (and Paddle Faster)
Let’s look at each in turn.
Just Own the Damn Robots (Part 2)
Back in 2017 Josh Brown wrote a piece called “Just own the damn robots” that reads like it was written today. In it he argued that the solution to a technologically-dominated future was to “just own the damn robots.” In other words, invest in the companies that are trying to replace you with software and machinery.
If you had taken this advice back in 2017, you would’ve done quite well for yourself over the next 7 years. And it’s quite possible that “just own the damn robots” is the right choice today as well.
But, before you go all-in on Nvidia or the Magnificent 7, consider the fact that this “solution” is already priced in. After all, why do you think these tech companies have outperformed the broader market since 2023? It’s because investors are already anticipating some sort of AGI future and are bidding up prices accordingly.
Of course, you may argue that they aren’t investing enough in these companies given your belief about the coming AI singularity, but that’s a separate issue. So if you believe that AGI is going to have us all replaced, then it wouldn’t hurt to own the damn robots.
However, I’m of the opinion that such a strategy may prove disastrous if our AGI future doesn’t turn out as planned. In the event we never get to AGI or it ends up being prohibitively expensive or a host of other things derail it, then you may lose a lot of money. As the phrase goes, “predictions are hard, especially about the future.”
Changing your allocation to be more tech heavy is just one solution to the post-AGI world. Another doesn’t involve your portfolio at all.
Hedge Your Labor, Not Your Portfolio
While the assets you own may be the perfect hedge to an AI-dominated future, what skills you decide to pursue may be even more important. Having capital will be important under AGI, but unless you are financially free, you will still need to work.
In this scenario, the ideal solution is to hedge your labor for a post-AGI world. What does this mean exactly? It means to build the skills that AI cannot so easily replace. What are these skills? Consider these:
- Taste & Curation
- One of the biggest problems with AI is its lack of taste. It can’t identify excellence because it doesn’t have a model for excellence. For example, give an LLM a great piece of writing, say from Morgan Housel, and ask it to “improve” the writing. If it had taste it would say something similar to what Pharrell Williams told Maggie Rogers, “I have zero notes for that.” Pharell has taste. He recognizes when greatness is difficult to improve upon. But AI doesn’t. Instead it will try to improve the writing while actually making it worse.
- If you want to go deep into a skill that AI won’t replace, develop a sense of taste and curation in your chosen field. Know when to recognize greatness and how you can use that skill to add value for others. This can be true with writing, design, food, and much more. AI will never replace Rick Rubin, so become your own Rick Rubin.
- Emotional Intelligence
- While chatbots are getting increasingly proficient at providing emotional support to humans, I doubt most people will want to hire them when they have to make emotional decisions. Sorry, I just don’t believe that chatbots are going to replace therapists, financial advisors, and similar roles. There is something about human-to-human connection and experience that you can’t train a model for.
- While some people may be okay with getting their advice from a highly advanced piece of software, I’m skeptical that this will be a realm where humans are willing to do so.
- Care and Comfort
- Imagine you have an aging parent who needs regular around the clock care. Are you going to have a robot do this or a human? Even if the robot is smarter than a human, I doubt you’d want to leave the fate of your loved ones to the hands of a machine. Don’t get me wrong, I can imagine how robots and AI will help the medical field in a host of ways, but at the end of the day I’m going to want a person there just in case.
- Get More Technical
- While there are legitimate concerns that AI will replace many of the “thinking” jobs in society, this is an even better reason to get more technical skills. If AGI/ASI does end up doing most of the intellectual work of the future, people will still be needed to manage and deploy this infrastructure. From AI researchers to data center technicians and everything in between, this transition is going to require lots of human labor.
- If I could go back in time, I would’ve been even more technical in my career than I am today. While I was able to make my technical skills more accessible (through charts and writing), if I was an even more talented coder this website would have many more calculators and other helpful tools. If you think the future is going to be run by robots, don’t beat them, join them.
At the end of the day, all of the skills listed above are my best guess about what could be valuable in a post-AGI world. Unfortunately, I know I will miss the mark with some of them. Predicting the future is so hard. If you had asked me five years ago whether AI would ever be able to create something that looks like art, I would’ve said no chance and I would’ve been wrong. Trying to understand how AI will integrate with our future civilization is something that we will all need to wait on.
If hedging your labor doesn’t seem like the right choice for a post-AGI world, there is a third option which may not require you to do anything at all.
Stay the Course (and Paddle Faster)
The last option to choose from in an AI-dominated future is to simply stay the course. If you own market-cap weighted index funds (like the S&P 500), then you will naturally increase your ownership of the tech giants as they get larger.
This is exactly what’s been happening for index investors over the past few years. They’ve been following a “just own the damn robots” strategy (to a lesser degree) whether they like it or not.
The thing that I support about this strategy is that it provides some allocation toward AI-facing companies without betting the farm on this single idea. Therefore, in the event that AGI doesn’t materialize as some hope, you shouldn’t see your portfolio crash as badly.
While staying the course is probably the most reasonable approach to take today, I can see why some might think it’s not enough. If you feel this way, then do what many in the AI community are doing now—work harder and acquire as many assets as you can before the advent of AGI.
This is a form of “stay the course” on overdrive that is also likely to provide the best risk-adjusted path forward. In other words, instead of changing directions, stay the course and paddle faster.
No matter what the future holds, it’s hard for me to argue against this strategy. I’m not one to promote hustle porn and working 24/7, but working a little more before a possible AGI future seems like a good bet. If you end up working more and we don’t hit AGI, no big deal. But, if we do get there, then you’ll be grateful you built a slightly larger financial buffer.
Whatever you decide to do, I wouldn’t worry about it because we have more time than we think.
We Have More Time Than We Think
While I don’t know a ton about AGI/ASI, the economist in me tells me that we have more time than we think. Here are a few reasons why:
- AGI won’t be cheap to build/operate. Once AGI/ASI is invented, it may turn out that it’s still cheaper to hire humans for certain kinds of tasks. OpenAI’s GPT-4 model reportedly cost $100 million to build and future models are expected to cost billions of dollars. Considering that OpenAI isn’t profitable at its current price point, you can see how there might be economic limitations on AGI’s development. And there may also be economic limitations for its use. The technology that makes ChatGPT possible is incredibly complex and that complexity costs money to build and maintain. Just like we built the Concord jet and then stopped using it for economic reasons, we may do the same with more advanced (and costly) AIs.
- Authors note: I originally wrote this post right before the release of China’s Deepseek open source AI model, which has taken the LLM world by storm. Based on what I’ve read, DeepSeek was much cheaper to build (~$10M) than prior AI models. I am not sure if this is true, but it demonstrates how quickly the AI space is evolving and why we won’t be able to predict its future with much accuracy.
- Intelligence is not experience. Even after we create superintelligent AI, it won’t be able to do everyone’s jobs overnight. It’s going to take considerable time to train and deploy these AGI/ASI agents to perform certain roles at a high level. This is the fundamental difference between intelligence and experience. Intelligence by itself isn’t enough to understand the ins and outs of esoteric systems and processes. We will need to train these models to do these things and this will take time.
- New technology creates new jobs. Every time a new technology comes around people warn of the coming end of jobs, but those fears never materialize. Yes, new technologies destroy jobs. Expedia displaced travel agents, Uber displaced taxi drivers, and smartphones displaced digital camera manufacturers. At the same time, we have more travel options, more private drivers, and more access to digital photos than ever before. The same thing will happen with AGI/ASI. This technology will kill some jobs while also creating new ways to work alongside these systems to be more productive. For this reason, the fear surrounding AGI is likely worse than AGI itself.
Don’t just take my word for it though, Sam Altman, OpenAI’s CEO, tweeted the following just a few weeks ago:
twitter hype is out of control again.
we are not gonna deploy AGI next month, nor have we built it.
we have some very cool stuff for you but pls chill and cut your expectations 100x!
If Altman is directionally right on AGI progress, then we don’t have much to worry about…for now.
Happy investing and thank you for reading!
If you liked this post, consider signing up for my newsletter.
This is post 436. Any code I have related to this post can be found here with the same numbering: https://github.com/nmaggiulli/of-dollars-and-data