Cory Doctorow Doesn't Really Understand Capitalism And Markets, Does He?

It's the confidence with which he doesn't that is so amusing

Apparently AI is just so over, Man:

Cory Doctorow

AI is the greatest money-wasting scheme humanity has ever invented

And, no, obviously. A rather better guess at greatest money waster ever is Mao’s China. Kept more people poorer for longer than any other one single human mistake. We could, I guess, think of the Soviets as being worse as they lasted longer, even if for fewer people, but. Such things as Peron insisting upon fascist economics just as Europe had proven it didn’t work are mere pikers by comparison with the big two there.

But:

What people don’t worry about is the fact that 35 per cent of the S&P 500 is tied up in seven firms, six of which have yet to make a penny’s profit on AI. (Nvidia, the seventh one, is profitable, because it’s the company the other six have given all their investors’ money to.)

AI then is the greatest money-wasting venture the human race has ever embarked upon. At $1.4tn in capital expenditure and counting, this industry – with an annual global turnover of less than $60bn – could become the most toxic investment bubble in history. As Stein’s Law has it, “anything that can’t go on forever eventually stops”.

Sure, this is a pretty big bubble but that $1.4 trillion? Pfft. The US has some $180 trillion or so of household wealth. Well under 1% is a rounding error. Bernie Sanders could kill more than that if he could ever get one of his bills passed through Congress.

Under 1% is also - not that I’ve checked, but I think I’m safe in saying - less than the ratio to GDP of Tulip Mania, or South Sea, or…..well, you get the picture.

But where the understanding really isn’t there:

Twenty-five years ago, I lived through the dotcom bubble. Many people point to that bubble and insist that even though that market frothed with silly firms, the underlying ideas were sound, and a bet on the web was a good one. It is argued that this means AI must be a good bet, too. This is an obvious fallacy: the fact that one thing stopped losing money and became profitable doesn’t imply that losing money is itself an indicator of long-term growth.

Ahhh, no, while the observation itself is correct it’s not correct in its usage.

Unlike AI, the web had brilliant unit economics. Adding a user to the web made the web more profitable. Every new AI user makes the AI sector lose more money. Each use of the web made the web more profitable. Every time you prompt an AI, the company supplying it loses money. Every generation of the web was more profitable. Every new generation of AI loses more money than the last one.

AI then is a normal technology, a grab bag of plug-ins that different people will find useful to different degrees. But it is also an abnormal bubble, vastly larger and more dangerous than the tech bubbles that preceded it.

No, canals, railways, cars, airlines and on and on have all gone through investment bubbles. And sure, unit economics is important. But it’s not important to the investment bubble itself.

The issue is that we don’t know. No one knows. No one knows which new tech - heck, whether there is even a new tech - is going to work. Work in that long term sense of being economic to run. Clearly, not all attempts are. So, we need to find out. Which is what the investment bubble does. Far too much - if you want to put it that way - capital flows into the field and everyone tries everything. After a sorting period we find out. We get the knowledge of what we did not know before. This works, or not.

Once we know whether it’s worth doing then we can start to optimise it. Sure, current AI is very brute force. But we’re seeing much lighter models (erm, China’s DeepSeek?) which provide 95% of capability at hugely lower operating costs. Once we know that AI works in the technical sense then that job of shaving costs - a la DeepSeek - will happen and we’ll find out whether it’s economically viable or not. My own projection - not that I have any technical knowledge at all - is that once we settle down on some standard AI algos then those are going to be done on optimised ASICs rather than GPUs. You know, just as happened with bitcoin mining.

Which is the bit Doctorow is missing here. It’s actually pretty normal that the capital spent upon finding out is pissed away. What matters though is that having found out can the tech be optimised so that it supports its own running costs? At which point it’s not so obvious that AI is going to fail, is it?

And even if it does, think on this for a moment. No modern railway has a hope in hell of covering its build costs as HS2 and the California debacle are showing so neatly. But all sorts of people still insist we should build ‘em……