My Son Won't Have a "Job" - And That's Okay: Embracing the AI Economy
My son is a computer kid. If he’d been born two decades earlier, I have no doubt he’d be living the FAANG dream right now—drenched in stock options, taking power naps in sleep pods, and strolling through gourmet cafeterias. He’s grown up watching me write code and solve problems, and as natural as it would be for him to follow in my footsteps, he can’t. When he graduates university in 2030, the world in which you could make a good living doing knowledge work will no longer exist.
We are on the cusp of achieving OpenAI's stated goal of making AI that "outperforms humans at most economically valuable work" and the level of concentrated disruption this will unleash on the economy and society is almost unimaginable. The traditional path—study hard, get a degree, secure a good job—will not survive the AI revolution. However, I believe humans can stay economically productive and meaningful even in this coming age of AI dominance. This isn't a paradox; it’s a vision of how the economy is going to evolve and the skills, mindset, and resources necessary to thrive in this new economic paradigm.
O3 Brings the End Game into Focus
On December 20, 2024, OpenAI unveiled its new o3 model and shattered the one barrier that might have slowed the future: operating beyond its training limits. For most of ChatGPT’s short history the sick burn in academic circles was to dismiss it as nothing more than a "stochastic parrot," meaning it could only engage in a probabilistic rearrangement of words from its training data without true understanding. But o3 has demonstrated that AI has broken free of the confines of pretraining and can now tackle novel problems never seen before by it nor any human in the public at large.
Unlike earlier generations of ChatGPT that would instantly fire off an answer retrieved from its memory (what scientists call System 1 Thinking), o3 is more thoughtful. Before answering it takes longer and first breaks down the problem into a series of steps. As it works through those steps, if an interim result doesn’t make sense, it might back up and try a different strategy. This newly developed capability to reason through problems (System 2 Thinking), has been particularly productive in math and engineering. On a mathematics benchmark intended to be challenging to PhD level mathematics professional, it solved 25.2% correctly. On a benchmark of real-world coding problems, it scored higher than OpenAI's chief scientist.
It just three months of training to drive the performance on the coding benchmark from a score of 52.9% to 75.7%. This rapid improvement was made possible because the new training paradigm no longer simply scrapes the internet in hopes of finding human generated examples of how to solve problems. Rather the o series of reasoning models, of which o3 is the latest, learns by reinforcing its own chain-of-thought reasoning when those reasoning steps led to correct answers. In other words, o3 doesn’t need human teachers. It can now teach itself.
Has AGI, Artificial General Intelligence, been achieved? The term, which broadly refers to machines with human level intelligence, is so ambiguous that it’s difficult to say. In many areas ChatGPT is superhuman. It’s fluent in every language, can write a better essay in seconds than most humans can in hours, and successfully fields millions of questions daily. The OG of thinking machine benchmarks was introduced by Alan Turing in his seminal 1950 paper titled "Computing Machinery and Intelligence". The Turing Test focused on language capabilities that were Sci-Fi impossibilities until ChatGPT blew past that level in 2022. A more modern attempt to benchmark intelligence, ARC-AGI, was developed by AI pioneer François Chollet in 2019. When o3 surpassed human performance on that benchmark, in typical fashion, the goal post was moved, and ARC-AGI-2 was announced.
Fabrizio Dell’Acqua described AI intelligence as a “jagged technological frontier” where some tasks are easily done by AI, while others, though seemingly similar in difficulty level, are outside the current capability. Psychological self-preservation has led society at large to focus on the ever-shrinking list of areas where AI falls short while glossing over glaring human deficiencies. However, this comforting period of self-delusion will soon end. Human capabilities are not improving. AI capabilities are on an explosive exponential.
The Inevitable Shift from Labor to Capital
It is generally acknowledged that AI will destroy some jobs. A recent Goldman Sachs report estimates “nearly two-thirds of current jobs are exposed to some degree of AI automation, and that generative AI could substitute up to one-fourth of current work”. But economic orthodoxy suggests worker displacement will be offset by new job creation and higher demand for existing jobs that are not automated away. Perhaps AI is just another in a long series of technological advances that have shifted the center of the job market: from working with our hands, to our heads, and maybe, in the future, to our hearts. As knowledge work wanes in the face of ubiquitous, commoditized machine intelligence, perhaps there will be an uptick in employment for waitresses, priests, yoga teachers, and other yet-to-be-invented automation-resistant occupations. As an introvert, career paths centered on human connection sound like a nightmare, but I imagine blue-collar workers felt similarly when told to "just learn to code" as a solution to technological unemployment in their industries.
However, it is entirely possible that this time is different. The population of horses and mules was at a peak of 25 million in 1920, but as these animals lost economic relevance that number fell continuously to a low of just 3 million in 1960. Will it be the same for human labor? Even if we don’t all get sent off to the glue factory, will the jobs that remain be capable of providing the employed the sense accomplishment and purpose that many of us currently feel with our career? For the type of person who reads an article like this, the answer is surely no.
Since the 1980s, labor's share of national income has been steadily decreasing, a trend often attributed to technological advancements that predate AI. Thus, it seems likely that the future economy will continue to shift in favor of capital. However, even given a likely shift of economic power to capital, there is a big difference in what society looks like if AI facilitates a centralized winner-take-all dynamic or if it supercharges millions of small businesses to compete in markets that were previously out of reach. In the answer to this question, I see reason for hope.
A Golden Age for Main Street?
A 2023 leaked internal memo from Google titled "We Have No Moat, And Neither Does OpenAI," laid bare the central challenge facing big businesses in the age of AGI. Instead of racing towards monopoly and infinite profits, the AI race is looking more like a race towards a commoditized bottom. While today it’s primarily the giants of OpenAI and Google that are pushing the frontiers of LLMs, open-source efforts at Meta and others are not far behind.
This commoditization isn't limited to AI labs; it's a fundamental shift affecting any business built on proprietary knowledge or processes. As AI becomes more accessible and capable of automating complex tasks, the specialized knowledge and skills that once justified large corporate structures will become increasingly less valuable. Effectively, any firm whose core value proposition is rooted in information asymmetry or labor-intensive processes is facing the same existential threat as the AI developers themselves.
In a world where AI empowers individuals and small businesses to compete in virtually any market, the only remaining "moat" for many will be location. Mass manufacturing increasingly will give way to hyper-local markets, similar to the diverse landscape of food and fashion. Everything from smartwatches to furniture will be customized to local tastes. These businesses, like many family enterprises today with generations pitching in—kids at the register, grandma in the kitchen—will be deeply intertwined with the lives and interests of their owners. The proprietor of a specialty fish breeding and coral shop, for instance, has a very different focus than the owner of a coffee shop—and that personal investment will be crucial for success.
In this AI-empowered landscape of small businesses, the traditional industrial-era framing of the economy as a class struggle between labor and capital becomes less relevant. In Q2 2024, FRED (Federal Reserve Economic Data) reported that 8% of national income fell into the category of “mixed income” or “proprietor’s income”—a category that blurs the lines between earnings from labor and returns on small business capital investments. This translates to over $2 trillion in economic activity – hardly a niche segment, and there is no reason to think that this category can’t grow in the coming AI era.
In a sense we may be going back to the future. For most of history, life was deeply local. If you needed shoes, you went to the cobbler; if you wanted bread, you bought it from the baker or made it yourself. The industrial revolution turned a simple product of flour, water, and yeast into a mass-produced commodity of bleached grains and dozens of additives under pressure from a need for manufacturability and shelf life. This sacrifice of quality and customization for scale is the story of the 20th century. With luck the 21st century will be a story of hundreds of millions of individuals and small businesses leveraging AGI to meet the specific needs and desires of local markets.
Planning to Win
When I talk about the future with my boys, I don’t lament the "tragedy" of them growing up in a world where they won't experience the "joy" of a 9-5 corporate job. I also certainly don’t let them confuse a future with less “jobs” as one with less opportunity to make money and thrive. Instead, I paint them an accurate picture of a future brimming with possibilities.
Someone born in 1910 started life in a world of Kaisers and Tsars, a world still largely powered by horses—a world that, in many ways, would have been recognizable to someone from a millennium earlier. Then, in a few tumultuous decades, everything changed. Empires crumbled, technology exploded, and the world was remade. But would any of us want to go backward?
In the same way, I expect there will be disruptions ahead of us. But alongside those disruptions will be amazing, mind-blowing, sci-fi possibilities. My boys aren't unlucky to be growing up in this era of upheaval. They get to be on the front lines of one of the greatest adventures in human history.