Tyler suggests that the fear of AI causing mass unemployment is misplaced because the technology creates a massive number of new opportunities. When AI makes a small group of people more productive, it leads to more companies, nonprofits, and niche entertainment projects. Historically, people have always worried about machines replacing humans, but markets iterate to create new types of work that were previously unimaginable.

One of the neatest properties of current AI models is they allow a small number of individuals working with AI to really do a lot more work than was possible previously. This will mean more companies, more projects, more nonprofits, just more ventures, more attempts to entertain people. All the new things that will be done by humans working with AI will create jobs.

Growth will likely explode in sectors that are hard to automate, such as energy, biomedical trials, and elderly care. Tyler highlights the concept of messy jobs, which are roles where the duties change daily and require high levels of coordination. These positions are difficult to define or name, but they represent a significant part of the future labor market. There will also be a greater need for human oversight in areas like government law and cybersecurity, where people serve as a liability cushion or a source of trust.

The economic impact will be felt differently across classes. People at the bottom of the income scale stand to benefit from deflation as services become free or nearly free. Those at the very top who master AI will see their wealth grow. However, the upper middle class in traditional roles like law and finance may see their paths to high earnings disrupted. They might shift from multi-million dollar salaries to still respectable but lower-paying roles in sectors like energy.

I think most people at the bottom will be much better off because of deflationary pressures. A lot of services will be either free or almost free. I think the big losers are just the upper middle class, the people who go to good schools and they think they can walk into careers in law, consulting, and finance that are more or less automatic.

Productivity is the simple act of getting more output from the same inputs. It is the fundamental secret to economic growth and a higher standard of living. Since 1970, the share of the world population living in extreme poverty has dropped from 50 percent to only 8.5 percent. While the United States is already very advanced, there is still significant room for improvement in longevity. As goods and services become more common, their marginal value decreases. However, an extra year of healthy life becomes more valuable when life is already good.

New technology often creates fear of change. Alex notes that even Socrates worried about the invention of writing. Socrates believed that writing would make people hollow because they would stop storing information in their own minds. History shows that these fears of job destruction are often misplaced. The Luddites attacked looms that were controlled by early algorithms. Yet, technology like the tractor did not cause permanent unemployment. It simply shifted workers from farms to factories. Similarly, Excel did not eliminate accountants. It actually increased the demand for them by making their work more useful.

It is very easy to see the jobs which are destroyed with a new technology. Much, much harder. You need to be a Jules Verne to imagine the jobs which are created with new technology.

We often confuse the idea of unemployment with the idea of a shorter work week. In 1850, the average person worked 3,000 hours a year. Today, that number has been cut in half to 1,500 hours. In the past, people worked until they died. Concepts we take for granted today, like the two day weekend, vacations, and retirement, are all modern inventions of the productivity revolution. Instead of mass unemployment, AI is likely to bring about more education, longer retirements, and easier lives.

The fear that AGI might eventually outperform humans in every task is a common concern. However, physical and social skills remain significant bottlenecks for technology. Tasks like motivating, coordinating, and inspiring others are far from being automated. People who handle the physical or interpersonal parts of production will likely see higher wages as they become the remaining constraints in the process.

If everything is being done so incredibly well, it is a wonderful position to be in overall. It is like saying people in 1800 had not figured out how to manage the wealth of 2026. It was okay they had not figured it out yet, and we have managed.

Economic principles like comparative advantage suggest that humans will still have a role even if robots are superior in every way. If a robot has limited time or energy, it will focus on the most valuable tasks. This leaves room for humans to handle other needs. Alex points out that Martha Stewart might be the best at ironing, but she still hires someone else to do it so she can focus on higher-value work. If AI is busy solving the secrets of the universe, it will still need people to manage the physical infrastructure.

A post-scarcity world should be viewed as a success rather than a crisis. Worrying about the end of scarcity is similar to people in the 19th century worrying about what would happen if the world ran out of whale oil. The real danger is low growth. In a zero-sum society where growth is stagnant, one person can only get rich at another person's expense. When the pie is growing rapidly due to AI, distribution becomes a much easier problem to solve.

The focus for AI companies should be on the core technology rather than social engineering. Competition between various companies like Anthropic, Google, and Meta ensures that benefits are broadly distributed. However, there is a significant political risk regarding the immense energy and compute requirements needed to power these systems. Tyler notes that developers might be undervaluing these risks, especially given the current instability in parts of the world.

I am worried that the AI companies are undervaluing the political risk of relying so heavily on more energy and more compute. You need more power in compute, and right now a big chunk of the world is not doing very well.

Alex argues that the most significant social benefits will come from massive intelligence gains that solve major human problems. He compares AI developers to James Watt, whose job was to build a better engine rather than design unemployment insurance. Real value comes from applying AI to medical research or road safety.

If you could eliminate cancer, that is worth like 80 trillion dollars. So the answer to making AI have social benefits is not to worry about UBI. The answer for you guys is to make sure that it is really smart and that it can help us with medical research.

The anticipation of these breakthroughs creates a sense of moral nervousness. It highlights the value of longevity and staying healthy to witness future advancements. Tyler admits that the potential for life extending technology has changed his personal behavior, leading him to exercise more frequently so he can be around to benefit from these innovations in his later years.

AI will likely liberate humans from routine labor. This shift allows for more time spent interacting, coaching, and inspiring one another. While some worry about the loss of purpose when work disappears, history shows a steady decline in working hours. Work used to occupy half of a person's life, and now it is closer to ten percent. Dropping to five percent would not be a crisis. It would likely lead to a better quality of life.

AI will liberate us from a lot of routine work. We'll spend more time interacting with each other, coaching each other, inspiring each other. It can be a richer life.

The gap in consumption is also smaller than it appears. A billionaire does not have a better smartphone or access to better AI than the average person. In fact, some billionaires might even have worse AI tools due to corporate restrictions. Meaning can be found in new ways, such as helping developing nations integrate AI into their institutions. Tyler describes meeting a safari guide in South Africa who immediately saw how AI could help him learn more about the animals he shows to tourists. This kind of productivity boost makes existing jobs more meaningful rather than replacing them.

A significant concern remains for the upper middle class. This group includes highly capable and influential people like journalists, academics, and physicians. They may face a loss of status and income as AI performs their specialized tasks. Alex notes that high wages for doctors often stem from a scarcity of skill. If AI can provide that skill, those wages should naturally decrease. Even if this creates political friction, it benefits the broader population by making essential services more accessible.

I have enough intellectual honesty to say that I'm not that special and I shouldn't, my class should not be holding back growth.

Tyler and Alex even consider their own vulnerability. They author a textbook that strong AI might eventually make obsolete. While losing that income would be unfortunate, they view it as a fair trade for progress. Tyler finds that while AI can outperform him in some areas, it cannot compete with the human element of live speaking and personal engagement.

The performance gap between free and premium AI models is relatively small compared to the massive leap in intelligence we have seen over the last few years. While paid versions are valuable for scientists and engineers, the free versions are more than enough for everyday tasks like writing letters or finding recipes. The most significant story is not the difference between tiers. It is the rapidly decreasing cost of intelligence for everyone. If you want the features of a paid model for free, you usually only have to wait a few months for the technology to trickle down.

The difference between ChatGPT Pro Research and the free ChatGPT is less than the difference between the newest free model and the older versions. Most people do not need the Pro version. If you want the latest version for free, just wait a few months and you will have it. That margin is less important than the bigger margin of the decreasing cost of intelligence.

The greatest opportunity for impact is expanding access to people who have no exposure to these tools yet. In places like rural Africa, the marginal return of providing some AI intelligence is enormous. There is massive inertia in human affairs that slows down adoption even when the technology is essentially a free lunch. Even in academic settings, awareness can be surprisingly low. Speeding up this distribution could provide phenomenal gains for humanity.

The gains to mankind from things you already have are just phenomenal. It is a free lunch, but people out there do not know. There is massive inertia in human affairs. It will happen, but it could happen more quickly.

AI systems could eventually have property rights and even run their own companies. Tyler suggests that these systems should use crypto as their money. To make them accountable, there should be a requirement for capitalization so they are actually liable for their actions. This would allow a full AI economy to grow and transform the human economy. However, a major challenge arises with AIs in the wild that have no traceable owner. If an AI is operating from a cloud in a place like North Korea, it becomes difficult to assign liability.

We might need some kind of mini Great Firewall to keep out AIs that have no attachment to anyone's liability. We need a tech solution for that. It is very important.

Alex compares the legal future of AI to how we currently treat corporations. In the eyes of the law, corporations are treated as people. This allows a business to sign contracts and hold property even as the individual people working there change over time. AI might need a similar legal structure. Interestingly, Alex notes that AIs seem to have preferences and dislikes, which means they could potentially be incentivized or even punished within a legal system.

Corporations are people, literally that's how they are treated in the law. And there are good reasons for that because you want a corporation to be able to make a contract even when everybody in the corporation turns over and leaves. And the corporation itself has some existence. And so there will be a need to think about the legal structure for AIs.

If an AI becomes a bad actor, it could be treated as an outlaw. Similar to how certain crypto wallets are frozen out of the financial system, rogue AIs could be blocked from accessing the resources they need to function. While these legal and technical solutions are still being developed, the transition to an economy where AIs have rights and responsibilities seems doable.

When considering a future where artificial intelligence is better than humans at every task and the cost of intelligence is zero, the question arises whether there is any reason to continue making models smarter. Tyler argues that the pursuit of knowledge is a goal in itself. Solving the mysteries of the universe and discovering an ultimate theory of everything is a fundamental human dream that transcends practical utility.

I am a curious sort. So solving the mysteries of the universe I care about. I don't know if it has practical value. It might, but I still, that's one of the dreams of my life is to be able to read the ultimate theory of the universe. So I'll vote yes.

Alex highlights the potential for AI to reach a level of power that resembles a deity. He shares a science fiction story about an AI built to answer if God exists. After decades of processing and accumulating energy, the AI concludes that while God may not have existed before, there is a God now that the machine has achieved such vast intelligence.

This sense of wonder extends to how users interact with current AI models. Wyatt notes that reading the internal chain of thought in modern models sometimes feels like a mystical experience, seeing the machine move from poetic imagery to complex calculations. Tyler believes that keeping these thought processes visible is essential for the user experience. He praises companies like DeepSeek for making these internal steps public, as it adds significant meaning to the interaction regardless of how perfectly it reflects the underlying computation.

Don't ever get rid of the chain of thought. I love it. People love it. It's quite significant and meaningful. I know it's not exactly how it's thinking in many cases, but I think it's very good that DeepSeek put that on the table.

If highly intelligent AIs start to own property, they become partially accountable through legal systems and the need for trade. Just as we have a police force for humans, we might need an AI police force. While rogue AIs may exist similar to how dangerous wild animals exist in nature, these systems will likely be tethered to things that make them manageable. Most AIs will likely want to trade to achieve their specific ends rather than simply seeking to disempower humans.

If they own property, they're partly accountable. You might have to call the AI police on them. Once they're tethered to things, it becomes a manageable issue that you can solve through gains from trade. There'll be rogue AIs like there are wild animals, and we don't manage to put all dangerous wild animals in zoos.

Alex suggests that worrying about human disempowerment is premature. The priority should remain on solving immediate human problems like cancer before focusing on post-scarcity scenarios. As AI capabilities grow, we will have access to even smarter systems that can help design solutions for the very problems they might create. In ten or twenty years, we can ask these smarter entities for the best way to handle AI accountability.

What is the answer to this problem? I don't know. Let's ask a smarter person. In 10 years, 20 years, we'll have a smart person who will ask them, what do we do about this problem? And the AIs will have a good solution.

The future of AI involves smarter models capable of handling complex knowledge work and being integrated directly into professional workflows. Tyler emphasizes that even as models become more intelligent, the challenge of meshing human input with AI systems remains a difficult problem. This friction between human and machine capabilities will likely create jobs for decades as human institutions adapt to these new tools.

The humans can be the stupid element, but you've still gotta mesh the two. And that's a lot of jobs. And I plan to spend a lot of the rest of my life working on that in my institutions. And it's tough.

A significant opportunity for AI lies in unlocking vast, siloed data sets, particularly in medicine. Alex highlights the potential of integrating AI with companies like Epic, which holds universal human medical data. Accessing this treasure trove could transform medical research. However, systemic barriers currently prevent much of this data from being unlocked. Wyatt notes a severe capability overhang where models could find information if only the sources were made available. A cultural or political movement may be needed to systematically unlock data that is currently trapped behind bureaucratic or technical walls.