Microsoft, Ozempic, and Progress
The original (Spanish) version of this article can be found here.
For months now, Silicon Valley has lost its mind. Several companies, fiercely competing for new markets, keep pumping out hyper-inflated headlines about the advent of so-called “artificial general intelligence” (AGI): a superintelligence supposedly about to replace, once and for all, all humans.
That doesn’t seem likely to happen. But in any case, it’s not even very clear what exactly an AGI would be. Would it be a machine capable of performing any intellectual task like a human? Would it be a universal intelligence far beyond anything human? Or would it be a conscious entity capable of experiencing its own subjectivity, understanding its own existence, and acting intentionally, even if it is not that good at solving problems?
This is no minor issue — if this technology really were to arrive, the future of humanity would be at stake in this discussion… and a lot of money too. Analysts expect that AI could attract up to one trillion euros in investment in the coming years.
Amid this commotion, a few days ago Microsoft and OpenAI — the company that dazzled the world with ChatGPT — took it upon themselves to produce a definition of AGI. One that actually explains what is happening in the world better than all the economics books on the shelf.
It turns out that Microsoft is OpenAI’s main investor: it has put $13 billion into the company. Despite all the shine, OpenAI is not profitable. It continues to lose money two years after launching ChatGPT and expects to keep losing money until 2029. Microsoft does not want to play the fool in this whole show, so it has secured its investment with a commitment: the startup will pay it 75% of its profits until OpenAI achieves the long-awaited AGI. When will that be? When the machine achieves consciousness? No — it will be when OpenAI reaches $100 billion in profits.
With this agreement, Microsoft and OpenAI have settled the debate and defined artificial intelligence in economic terms: they say we will call it AGI when it has generated so much profit that it is clear it can replace humans in the productive system. So no consciousness, no comparison with the human mind, no Turing test, nothing. There is no need to get lost in complex philosophical discussions when what we came here to talk about is money.
This summit pact between tech giants is the perfect example of what is perhaps the main problem of our society today: we do not know how to measure progress unless it generates money.
Of course, up until this millennium that was not a problem, because material progress and human progress went hand in hand. Even when women began to achieve higher levels of equality, GDP went up (a lot). But this has not been the case for years. For a reason that has to do with the nature of technology, and another that has to do with the nature of the economy.
When progress shrinks GDP
I would say not only is it possible, but it is most likely that generative AI will transform many things without producing a substantial direct growth in the economy. This is what happened with email, with instant messaging, with Google search, with GPS, with database software, with cryptography, with digital photography, with open publishing technologies (blogs and social networks), or with voice recognition technology, among another million things. For a well-known reason: when a technology is developed, as is happening now, in a distributed manner across several universities, companies, and research institutes, it is not possible to “keep” the patent to be the only company offering it to customers and control the price. As soon as the required knowledge is born, many companies start competing for the same segment, and prices plummet. This is called “commoditization,” and while it also happens with physical products, it happens much sooner and much faster with digital products.
To give a simple example: this is what happened with mobile phones when the technology they used went from being only in the hands of Apple and Samsung to being practically ubiquitous. Then, hundreds of brands started producing devices, and prices collapsed. Today, phones are a commodity. They are all practically the same and cost just a handful of euros.
Another paradigmatic example is that of solar panels, which in 50 years have gone from costing $130 per watt of power to $0.32. That’s about 500 times less.
This is exactly what is happening right now with “AI.” After ChatGPT, other bots appeared that are today indistinguishable from the first. And while it is still hard to find real uses for this technology in the economy, it has already spread so widely that it has become ubiquitous at very low prices.
But of course, that doesn’t mean it won’t change the world — only that it won’t generate the mountains of money expected in Silicon Valley. On the contrary, more and more technologies are producing so much value and progress that they actually wipe out large segments of employment and render entire industries obsolete, as has happened with Novo Nordisk.
Novo Nordisk, the manufacturer of the revolutionary weight-loss drug Ozempic, lost a staggering 90 billion euros on the stock market last week. Analysts say the reason is that a clinical trial on one of its drugs did not deliver the 25% weight loss shareholders expected (which is huge), but instead 20% (which is still huge). But behind this stock crash lies a bigger worry: investors are concerned about the development of another 40 similar drugs by other companies, about to hit the market. If Novo Nordisk cannot produce another much more revolutionary drug (that makes you lose, I don’t know, 100% of your weight? :p), it will have to lower prices, and by a lot. Ozempic has been on the market for seven years and only four as an anti-obesity drug, but in that short time it has turned the world upside down. It is the commoditization of miracles.
What happens in macroeconomics when a miracle drug ends the world’s leading health epidemic, affecting one in every eight people? Does this immense leap of progress get reflected in global accounting? No — the opposite happens, GDP shrinks. It goes like this:
At first, very few users buy the medication at a high price from the only manufacturer holding the patent (Ozempic costs several hundred euros per month and is expected to be used for life).
Attracted by the chance of making profits, all the pharmaceutical companies rush to launch their own drug. Within months the market is flooded with Ozempic copies and prices plummet. There is not much problem — except the crash of Novo Nordisk — because the market is gigantic, and many more people become users.
Millions of people lose a crazy 25% of their body weight. Global obesity is reduced by, say, 50%. All those people no longer need to go to the hospital for obesity-related conditions. It is estimated the world spends about 800 billion euros a year just on treating one of those, type II diabetes. Global healthcare spending decreases substantially.
Sales of diets, personal trainers, nutritionists, gym memberships, endocrinologists, and diet products decline.
The result is less economic activity, fewer jobs, and fewer money transactions as a consequence… of progress!
The usual refrain is that these phenomena create opportunities for new industries. For example, if people are no longer overweight, they exercise much more, and sales of sportswear skyrocket, or that if you can eat whatever you want without fear of gaining weight, food sales shoot up.
But what has been happening for decades is that these new industries are just as affected by the same phenomenon of “commoditization” as the previous ones, and we get into a spiral where that expected economic growth never really materializes.
And, in any case, it makes no sense to have a system that measures GDP losses, but does not have another metric to measure the value of ending obesity.
The Crocodile Economy
This has been happening, really, since the dawn of economics. That measurement system based on things having a price is like a crocodile. Its eyes are on top of its head, and it is only able to see a limited part of reality. Everything that happens below the water’s surface escapes its notice. That is why it has always had trouble understanding and capturing things that were very valuable but were not exchanged in a market.
For example, it is well known that the economy does not measure the unpaid work that families do in the home. This is not exactly a rounding error: according to the International Labour Organization, these hours represent 25% of all work hours, and if they were paid, they would be equivalent to 9% of global GDP every year.
It is also known that the economy does not measure the effects that the consumption of natural resources has on the environment, which some authors estimate to be worth another 5% to even 20% of world GDP each year.
And there are other, lesser-known but no less paradigmatic examples of this blindness. For instance, the benefits that citizens receive are accounted for differently depending on whether they come from the public or private sector. If they come from the private sector, they are counted at their price, just like any other product, but if they come from the public sector, they are accounted for by the cost they represent to the state.
According to this criterion, the American healthcare system (which represents 17% of GDP because it is private) should be better than Spain’s or France’s (7% and 10.5% of GDP, respectively), when it is obviously not. So outdated is this way of measuring value that if the United States had universal public healthcare, its GDP would fall by several points.
The same happens with private education or with long-term care for dependents. A country with a smaller public sector will have a higher GDP.
If we add up all these magnitudes, which are by no means the only ones, it is unthinkable to argue that this is merely a measurement problem. It is not, as some economists have suggested, a sort of irrelevant “blind spot” that econometrics cannot reach. The crocodile metaphor is much more accurate: it only sees what interests it — the animals floating on the surface of the water. It is an accounting system for just one part of human exchanges. It only measures things when they enter or leave its area of measurement.
In fact, the growth of the entire industrial era can be explained precisely through this mechanism. Between the 18th century and the late 20th century, an endless list of processes moved from the private sphere (which the economy does not measure) to the markets: Clothes stopped being made at home and a fashion industry was born. Children stopped being raised entirely within the family and nurseries and preschools were created. Education took youth training out of the hands of guilds, elder care shifted to nursing homes, clothes stopped being washed by hand and millions of washing machines were sold, heating systems were installed in homes that depended on fuels you could not gather (like wood) but had to buy on the market. Families stopped growing their own gardens and baking their own bread. With each task that left the household, demand for jobs was created in factories, shops, and public services.
To produce all these things, natural resources were used that had never been included in accounting books until then. Mines, oil fields, arable land, and, above all, land in large cities were incorporated into each country’s national accounts. Today, two-thirds of total world wealth is “stored” in that urban land. Nothing to scoff at.
In this way, the reason why the economy “grew” over the last 250 years can be explained by the fact that it was gradually incorporating things it did not measure before, but now does. As if a crocodile thought the lake was full of fish because they had all come to the surface to feed.
And what has been happening in recent years? Precisely the opposite. As we saw in the Ozempic example, technology is increasingly produced in a distributed way and becomes available to anyone very quickly. The only reason we won’t see a generic drug with the same formula as Ozempic in the coming months is that this particular formulation is patented — but there will be others.
It was the same with social networks, with email and blogs, with Internet search, with the web, with vaccines, with DNA sequencing techniques, and with many other things. Technologies disseminated so widely that they had a gigantic impact on society, but probably caused a reduction in that portion of the world that the crocodile economy sees. The same is going to happen with solar energy, which, thanks to plummeting prices, will mark a before and after in human history.
More and more, technology extracts things from that part of reality perceived by the crocodile. That does not mean it does nothing — far from it — but rather that reality, change, and progress happen beneath the water: they cannot be measured in money.
That is why it is urgent to think about other ways of seeing the economy that do not only measure what is exchanged for money. Or, even better, to invent a new science that does not dedicate itself to the “allocation of scarcity,” as economics does, but that aims to study all the things that are abundant and exchanged without money in between. And that measures how much we have progressed in recent years and how much we can continue to progress in the future in so many fields that are undervalued today.
From the Greek word plutos (wealth, fortune, abundance) and nomos (to distribute, to allocate), we might call it a plutonomy: the study and allocation of abundance.
To be continued…