Is Capitalism Bad for Technology? Or Fund Government Science

More of a musing than a post, today. Life has been hectic, unfortunately, and so I didn’t have the time to research a longer post. But this has been something I have been mulling for a few years, and the deeper we get into the imitative AI era, the more I am convinced this is true.

Imitative AI is not a transformative technology. This is not to say that it has no uses — it clearly does, especially in low-quality domains (early customer service and boilerplate content creation like marketing copy and simple business emails and presentations) or domains where internal consistency is more important than external validation (like programming). But it is not going to transform really anything. Even in domains where it is best suited, it doesn’t live up to the hype, and the productivity gains are often questionable at best when they aren’t illusionary. And that is even before we get to the deskilling issues. All of that would not matter if the systems were not such giant money hogs. It is inconceivable that any firm is going to make money on imitative AI in the near future given the combination of cost and middling results. It is something of a dead end, at least now, more fit for experimentation and research than business applications.

And yet, we see it pushed into every aspect of our lives. Companies use it as an excuse to fire people and insist that it must be everywhere if we are not to lose out on the greatest invention since either sliced bread or the entire industrial revolution. Why is such a poor business being pushed by so many businesses? Capitalism.

In a sane world, imitative AI would be relegated to universities or the research departments of some larger firms with a bit of cash to burn. It should not be the central focus of the American economy, and it should not be hyped as the be-all and end-all of all industries. Any sane business would be focused on electrification or climate resiliency or mRNA vaccines or machine learning-driven automation. That is where the money is likely to be, given the actual challenges the world faces, but all of our money seems to be tied up in the probabilistic Clippy machines. The incentives that capitalism produces are completely broken and push us toward waste rather than a solid economy.

All of the things I mentioned above, with the exception of machine learning automation, are capital-intensive. You can add things like medical devices to that list, if you want, another likely expanding market, but the difference between imitative AI and these others is that the AI alternatives require investments in physical goods to come to fruition. And those investments mean that they likely cannot have the same theoretical returns as purely software, or almost purely software, based products. And that is BAD to the people who run the economy — Wall Street and venture capitalists — since they depend on constant growth to keep making the money they believe they are entitled to. Far better, to them, to bet on a long shot in the hopes of it paying off than to plug away at medium shots with much better success chances but lower theoretical payouts. In their world, the incentives are to get rich quick, not to build a sustainable economy that helps society. And those incentives are made even worse by the tendency toward monopoly.

The end result of capitalism is monopoly or near monopoly. Competition has winners and losers, and at some point, the winners get too big to compete against due to their structural advantages. That is why anti-monopoly rules are the most essential ingredient in a capitalist economy. Without a smack on the nose with a rolled-up newspaper every once in a while, businesses will pee all over the economic carpet, so to speak. Unfortunately, we have forgotten that basic rule and allowed monopolies, especially in software-connected industries, to flourish. In a normal-functioning economy, a firm like OpenAI would likely be dead by now. It is not a viable business, but it was bailed out by Microsoft at least once. Microsoft, Meta, and Google can keep these firms running because they have monopoly rents that they need to find ways to invest. And in this case, they are looking for things that will please Wall Street — steady, large growth. With those incentives, it makes sense that Microsoft would prop up OpenAI’s corpse in the hopes that they will get a payoff. Easier, surely, than forcing Wall Street to understand the concept of a mature business or finding things that customers really want. As a result, we have an immature, possibly dead-end technology driving our economic fortunes. Capitalism has created, predictably, a potential disaster.

I am not, by the way, saying that a group of people who look nuts to the rest of us cannot create a viable business. The passenger airline industry looked insane after World War II for a variety of reasons, and yet we have passenger airline travel (whether it is a viable business is a question for another day, but it’s viable enough to keep planes in the air). IBM would never have invented personal computing, and the mRNA vaccines got people nearly fired. But there is a difference between people taking a bet on something they believe in and incentivizing a system where wrong bets take down a significant chunk of the economy and do real harm to people in the meantime. We now have the latter, and we should be taking steps to prevent it. Imitative AI is a fine subject for university research (where a lot of the material that drives our economy, up to and including the internet itself, came from), but we shouldn’t have a system that rewards fantasies over reality.

But, but, but, I can hear the cries, incentives! Without the possibility of getting rich, no one would have ever created any businesses, and we would have lost all those advances. If you believe that, then you are an idiot or an economist (for a field that has so many smart people it has more blind spots than a buried semi-truck). Economic incentives are not the only incentives people respond to. In fact, economic incentives distort human behavior much more than they reveal it. People who need not worry about their material needs would lead much different lives. If you were to put a salary cap on the United States, with minimum and maximum salaries for everyone relative to GDP, you would lose not a moment of invention, not an instance of creativity. People invent and create because that is what humans do — we are curious little monkeys and we stick our ignorant paws into pretty much everything we can get them into.

Take this little newsletter. I have less than five hundred subscribers, far too few to ever make any money off this, as much as I love all my readers. And yet, here I am, pouring hundreds of words into this each week. Same with my failed fiction writing career. I write not in the same vain hope that I will become a millionaire but out of the pleasure the process gives me. Doctors treat patients and do research because they want to help people. People create art because they want to connect with other people. Scientists and programmers research AI because it’s a fascinating subject. Almost no one would stop that work because they couldn’t be Zuckerberg. In fact, you almost certainly get more art and research and cool inventions if more people had the space to not worry about being homeless and hungry. Economics limits opportunities, and capitalism creates incentives that hurt rather than help people.

This ended up being longer and more rambling than I originally thought, but, hey, that’s the fun of it — for me at least. If there is a takeaway, I suppose it would be this: the current system is not the best we can do. Capitalism, especially as practiced in America, actually retards human progress and well-being far more than it advances them. Imitative AI is just another canary in that particular coal mine. Maybe this time, we should pay attention as it chokes and try and find a better place, a better way, to live.