Addictive machines

Computing machines are addictive. Internet and gaming addictions are more or less recognized phenomena, but I don’t know whether we talk about “computer addiction” as a more general phenomenon. In classical behavioralist terms, it’s periodic reinforcement: occasional, unpredictable positive stimuli that result from using the machine—as you use the machine, it occasionally rewards you with a sudden feeling of success that encourages you to use it more. This can be easily compared to gambling addiction: it’s not very far from a slot machine to a computing machine. Like the slot machine, the computing machine provides and addictive feedback structure that: it periodically rewards our efforts just often enough to condition us to continue poking at it. Ask a programmer, and they’ll tell you that they do the work because it’s addicting. There’s a reason we feel such a drive to debug code: we’re chasing the high we get when it finally works.

Now, some people profit from making their applications particularly addicting: youtube, tiktok, and facebook make their daily bread by addicting the user, because they need the user to use their product in order to profit. In Yannis Varoufakis’s recent interview with Evgeney Morozov in The Crypto Syllabus, the former finance minister of Greece and self-described “erratic Marxist” explains how this works. “Command capital,” he says,

lives on privately owned networks/platforms and has the potential to command those who do not own it to do two things: Train the machines/algorithms on which it lives to (A) direct our consumption patterns; and (B) directly manufacture even more command capital on behalf of its owners (e.g. posting stuff on Facebook, a form of labour de-commodification).

In other words,

Any digital commodity has command value to the extent that its buyer can use it to convert expressive everyday human activity into the capacity to train an algorithm to do two things: (A) make us buy stuff, and (B) make us produce command capital for free and for their benefit.

That is, the “product” sold by the likes of Jeff Bezos, Mark Zuckerberg, and the boys at Google is their ability to modify, or control, the users’ behavior. They acquire this power to control by getting people to use their service: the more you use it, the more they know about you and the better able they are to control your actions. And the best way to get people to use the product is to make it addictive.

This is something like the cigarette company’s strategy with tobacco: they didn’t create nicotine, nor the tobacco plant it’s found in. But once they understood its effects, they did their best to maximize how addicting their product was so that the customers would be driven to buy more. The strategy at the tech firms is slightly less direct—rather than selling the product to the addict they sell the control they have over the addict—but their strategy is the same: profit off of addiction. The key difference is that the addict of the tech firm doesn’t pay anything; they don’t have to: their addiction pays for itself. The only cost is that they are controlled by whoever bids on the right to control their behavior.