Computational Tyranny

20 Jul, 2025

We are under constant cognitive assault.

Buying plane tickets is navigating a minefield - one misclick blows a hole in your wallet.

Resolving a mistake on your utility bill is a tactical operation: dodge hold music traps, outwit the chatbot, then convince the agent you shouldn’t need to pay your neighbour’s bill.

Our decisions aren’t shaped by direct threats or outright lies. Governments and corporations work together to overwhelm our cognition and they benefit from our weakened state.

Everyday activities require more and more thinking. That's bad. And computers made it worse.

That’s computational tyranny.

The tyranny distilled

Computational tyranny is the rapidly increasing complexity in all areas of life where we bear the costs (in time, money and sanity) and where we are exposed to large risks. More institutions impose complexity on us, we barely have a say in what the rules are and there’s no incentive to simplify. All accelerated by cheap computation.

That’s the gist, let’s talk details.

The complexity burden falls on us

The tax authority sent me a letter. Apparently I miscalculated my taxes and I owe more than I already paid. They want their money back.

The problem is, they are wrong. The tax man misapplied their own rules and now I need to hire an accountant, dig up my old records and make a case for why they are wrong. It costs me tens of hours of my life and hundreds of pounds in costs. The letter I received was one of a couple of thousands sent by their automated system. It costs them almost nothing and if someone is not able to respond, they get my money.

That’s complexity cost asymmetry. The government created a complex tax system and there’s no simplification in sight. The cost is horrendously asymmetric: the tax authority amortises the cost across all taxpayers but each taxpayer has to spend their time and money to resolve any issue which comes up.

Even worse, risks are also skewed against the individuals: the tax man faces no penalties for wrongly accusing someone of underpaying tax whereas the individual might face substantial fine or even jailtime.

And there’s nothing keeping the complexity in check. There are virtually no downsides for the tax man for imposing larger and larger cognitive burden on the taxpayers. The burden is already high. If HMRC asks me to pay £500 or else, I’m simply going to pay. It’s cheaper that way.

What’s bearable about the tax authority story is that you aren’t going to be in this unfortunate situation often. You have to bear the cost but it’s likely you’ll find the time to deal with it.

Unfortunately, it’s not just the government increasing the daily cognitive burden. The assault comes from all directions and no area of life is safe.

Everyone wants a piece of your brain

Say you are in a market for wireless headphones. You search for "best wireless headphones 2025" and you are presented with suspiciously similar lists from suspiciously similar websites as far as eyes can see.

It’s not your first time on the internet so you added "site:reddit.com" to your query. A lot of accounts with 4 numbers at the end of their username seem to be recommending the same pair of headphones. That doesn't look right. Let's try comparing the technical parameters.

They aren't easy to find and each manufacturer uses their own terminology for the same or very similar concepts. The measurements they tell you about are not exactly comparable. Why is it all so complicated?

That’s three hours wasted, three hours of your life you are not getting back. You buy the pretty blue ones you saw on Instagram instead.

It’s nothing new. The internet advertising warfare has been going on for a while. What's new is that people don't have the time to spend on making the right decision. Getting necessary information became more expensive and the amount of thinking capability left in us after dealing with demands of everyday living is low.

Everyone wants a piece of your brain, it's getting fried, and they're monetising your bad decisions. The incentives are clear: keep your cognition low and we'll get paid more.

Hello? Is there anybody out there?

Firstly, computers help organisations to avoid contact with individuals. We had the long phone line queues for a while (“we are experiencing an unusually high call volume at this time”), it’s an old trick. Now, we have AI-backed chatbots. No longer there’s a “Speak with a human being” button prominently displayed on the website. You have to play a game first: convince the chatbot that your issue is important enough to speak with a person.

The more advanced the AI-backed chatbots get, the more plausible deniability for institutions. Bots will help with a large volume of simple queries, the time-to-resolution statistics will look amazing. As long as you exclude the people who never managed to get the help they needed, of course. An infinite time-to-resolution doesn’t look pretty in the quarterly report.

Secondly, the person you finally get in touch with can’t do much for you. They are bound by written procedures they are not allowed to deviate from. They have limited views of the company they work at and they can’t understand your issue deeply enough to help you.

It’s the irony of automation. The complexity of the internal procedures in institutions is becoming so high they are inscrutable even to people designing and implementing the rules. Even the most kind-hearted person wishing to help you will not be able to: they don’t know how and they can’t know how.

You have two options. Option one: give up, and bear the cost (e.g. pay the extra money on your bill even though you know it’s fake). Option two: lawyer up.

Unreasonable computers

In most democratic societies the ultimate recourse is going to court.

If the other party is unreasonable, like your utility company claiming you are responsible for a neighbour's utility bill, at some point you need to sue. And suing is expensive.

It takes a lot of your time, lawyers cost a lot of money and there’s the debilitating uncertainty of whether you are actually going to win your case. Remember, being right is not what matters in court.

How are computers to blame here? The core of the problem is that the algorithm is free but the lawyer costs £250/hour.

Computers make human time more expensive. It’s about a Baumol’s cost disease style argument. The more we automate, the more a single person can achieve in an hour, the more expensive human time is.

Computers also help organisations be unreasonable at scale.

Decades ago, if an organisation wanted to issue you a bill, they would have to have an employee type out the letter, address it and send it. If the bill was issued in error, they would need to have someone staff the phone line and spend time resolving the issue. The costs weren’t large, but they weren’t negligible.

Not so much with computers. You ask your analyst to write a SQL query, automatically generate thousands of emails and off they go. Support costs are also much lower, most people give up after waiting for an hour on the phone, only to be greeted with a maze of options, all leading to “chat with us”. And the chat is obviously a shitty AI agent whose only goal is to keep you away from a human as long as possible.

What’s worse, the capability to be unreasonable at scale is available to most institutions, organisations and corporations you interact with. It’s the same with your university, mobile provider, utility companies, local government, your bank, your credit card processor and so on. You need to deal with more entities and they each can make your life hell - at zero marginal cost.

Nobody’s fault

You get rejected for insurance, seemingly for no particular reason. Or you can’t open a bank account and you’re told you “are not allowed to ask about the reasons”. Who’s at fault?

Computers don’t make mistakes. Algorithms are free from bias. These are common lies which allow companies and governments to avoid responsibility. It’s a rehashing of the old favourite, “these are the rules, I can’t help you” but more insidious: it’s harder to prove malice the more complicated and opaque the system is.

Responsibility laundering is a time-honoured technique, employed by governments and corporations worldwide. It’s easier with computers although it still requires a bit of skill to pull off: you can’t just email your engineers to “make the algorithm reject these people”.

Still, happy accidents happen. An algorithm is deployed and it just happens to be slightly better than the previous one, it refuses to give loans to these people slightly more often than the previous one. I’m sorry, I mean it increased the Q3 revenue, so we continue to develop the algorithm in the same direction.

There are efforts underway, particularly in the European Union, to make sure people are not at the algorithm's mercy. While this is feeling my heart with optimism, it’s not enough.

Meanwhile the revenue keeps increasing, individuals are getting screwed and nobody’s at fault.

Where does that leave us?

The problems I described aren’t new. They are partially the classic capitalism critique, partially a badly plagiarised Kafka story. Still, a new phrase tying the computation to these issues is needed.

A large quantitative change is indistinguishable from a qualitative change. That’s what happens when cheap computation meets Western society: it functioned with a background assumption that people won’t create more complexity than they can handle.

Computers changed that and bigger players like corporations and governments gained a big systemic advantage over individuals: they make the rules but they don’t bear the costs of the rules breaking. They increase the complexity of everyday life and instead of being punished for it, they are rewarded for it.

Previously, you had to spend human time each time you wasted another person's time. Maybe less time, like a couple of minutes to draft and send a letter versus a couple of days for the recipient to deal with it. Still, it used to put an upper bound on each organisation's time-wasting potential. There’s no longer such a bound: the marginal cost of wasting an hour of human life is zero.

What bothers me the most is that the biggest cost of this madness isn’t economical but moral: we are wasting people’s lives for no good reason.

Does wasting 80-human-years feel that much different than ending someone’s life? Are we willing to keep squandering the only truly limited thing, our time here, on negative-sum games benefiting a few? Why are we collectively fine with the tyranny around us?

Ideally, I’d follow with a “Solutions” section here. It turns out that noticing a problem does not imply you have any half-decent ideas about fixing the problem.

I want people to be aware and have a name for the issues described above. The solution would likely have to put some, if not all, of the complexity burden on the institutions imposing it on everyone. That’d be a start.