The Cybercrime Economy and the Inverse Singularity

There is a really boring invisible plague consuming the world secretly contributing to poverty, starvation, and death. It is a parasitic evil empire draining a massive amount of the wealth and prosperity you will create throughout your decades of work providing goods and services to others.

This evil empire is the cybercrime economy. Warren Buffett has correctly said cyberattacks are a bigger threat to humanity than nuclear weapons—calling them "the number one problem with mankind."

The size of this theft from the productive economy by the parasitic cybercrime economy is greater than the total government spending of every country. This theft is greater than the entire economies of every country in the world except for the US and China.

If it's really such a big deal why am I never hearing about this?

Why don't I hear more about cybercrime or notice this massive theft if it's such a big deal?

It's boring. We evolved in a world without computers to be worried about violent threats, not abstract economic parasites. You can't take a picture of cybercrime on the newspaper and inspire fear.
Victims hide their attacks. Revealing these attacks opened them up to further lawsuits or reputation destruction.
The costs are hidden. Like inflation, they're mixed in to the cost of everything that you buy. Companies have to pay insurance or ransom as a cost of doing business and this is passed on to consumers so no one knows how much cheaper things would be without the cost of cybercrime.

The cost is $10 trillion divided by 8 billion people. The average annual cost of cybercrime per person would be approximately $1,347 in 2025, assuming a global population of 8 billion people. This is money that could have been spent on curing disease, feeding people, or creating innovation—instead it's stolen by parasites.

Cybercrime often goes unnoticed or underreported due to fear of repercussions, difficulty in detection, lack of awareness, shame, reputational risks, and internal disclosure concerns within organizations. These factors contribute to a lack of public awareness regarding the true extent and impact of cybercrime.

You may say, "I've never lost any money from cybercrime. That just affects big companies." You're wrong. These costs are just passed on to you through:

Price increases
Lack of innovation
Less economic growth

Why does economic growth matter? Because it ends poverty and represents the creation of goods and services that solve human problems. So, if you don't like getting diseases and dying, it's going to take a lot of resources to cure the over 10,000 currently uncured diseases. These resources come from economic growth.

Key Facts and Statistics

The cybercrime economy is the 3rd largest economy in the world, after the US and China, and it's growing exponentially. In 2021, the global cost of cybercrime reached $6 trillion, and it's estimated to reach $10.5 trillion by 2025 (Cybersecurity Ventures).
Small businesses bear the brunt. Extortion or ransomware shows up in 88% of small-business breaches, versus 39% at large organizations (Verizon Data Breach Investigations Report)—and small firms are the least able to absorb the hit. Consequences:
- Crushing innovation: most innovation comes from small new startups
- Monopolization: only mega-corporations can afford this cost of doing business
Criminals are rarely caught—and cybercriminals almost never are. Even murder is solved only about half the time, and ordinary theft usually goes unpunished. Cybercrime—theft or extortion committed from across the world—goes unpunished in the overwhelming majority of cases.
One sentence of 20 years was recently handed down—among the longest on record for this type of crime.

Simple math shows that soon the world could be overrun by millions of AI parasites feeding on the productive members of society. Given the system we have, the coldly rational move for the smartest, most resource-driven minds is to become a cybercriminal: there's a near-zero chance of being caught and almost unlimited upside compared to a normal job.

The Inverse Singularity

For almost all of human history, average living standards were basically flat. Then after the Industrial Revolution (and the advent of capitalism) we got a system that rewarded people for cooperating and producing, with a decentralized set of incentives that reward collaboration even when nobody understands the whole picture.

That's how you get things like a pencil for $0.10, even though it took hundreds of thousands of people to make it. Most of them have no idea what the other people in that supply chain are doing. The same is true for basically everything: the chainsaw, the graphite, the shipping, the factories, etc.

Then we got democracy and stronger protection of individual property rights, so people could keep the rewards for their labor instead of having it stolen by the government. We also got better protection from criminals and theft, which further incentivized labor, innovation, and long-term investment.

All of that drove huge improvements in human health and standard of living, and it can be accelerated further by artificial intelligence.

If AI can eventually do more of our work for us, and also improve itself, then it will get better at improving itself. That means it gains capability faster and faster, becomes more efficient, and can develop solutions to problems that currently block human progress, survival, health, and happiness.

But that requires the initial conditions to be set so that counterproductive activities (like extortion) are costly, and productive innovation is rewarded.

Right now, we're near the onset of a possible explosion in AI capability, but our incentive system has a much greater reward for harming the producers of goods and services than it does for doing positive things. For example: a hacker/cybercriminal can potentially make millions of dollars a year doing theft and ransomware, versus someone doing positive work for society, who might make 50 times less.

The Singularity Analogy

The clearest thing in nature that represents a singularity is a black hole. As you approach it, the gravity increases faster and faster. If you graph "distance to the black hole" on the x-axis and "gravitational force" on the y-axis, it looks like a hockey stick: at first it rises slowly, then it shoots up toward infinity.

The "technological singularity" is an analogy. As technology improves over time, each big invention makes future inventions easier and faster. Progress compounds. If you eventually get AI that can improve itself, then each improvement makes it better at improving itself, and that feedback loop accelerates.

You can already see the early version of this: most people aren't aware of current AI capabilities, partly because things are moving fast and people have limited time and attention. Even a year ago, many statements about what AI "can't do" seemed true, but recent developments have surprised even many of the people building AI.

Why the Initial Conditions Matter

If the AI's initial conditions (its goals and incentives) are set toward pro-social outcomes (curing diseases, preventing starvation, keeping people housed and safe, stopping violence) then a singularity could be incredible.

But if we don't set those conditions, then we could get the opposite.

You can create regulations to try to stop AI, but if you succeed, it's more likely you stop good actors, and you slow down the development of beneficial AI (the kind that could cure diseases, etc.). It is not going to stop criminals or states like North Korea from training models and using them for hacking.

Why AI Makes Cybercrime Scale

Right now it's expensive to "hire" another cybercriminal. There's a long gestational period: it takes ~18 years to go from a baby to a full-grown adult who can execute attacks. And if you're a programmer, you might make like $200,000 a year.

But soon, attackers will be able to train little AI agents (little electronic people in the computer) that can do what humans do now. That could decrease the cost of each "employee" by maybe 1000 times. So they can increase the number of targets they hack by 1000 times.

Right now, for it to be profitable, targets usually have to be rich: big companies that can pay a lot. But if the cost of an attack drops by 1000x, it becomes profitable to target everyone.

They can go find pictures of you and audio recordings of your voice. They can create videos of you saying anything or doing anything. They can hack your email, train it on things you've said, and make it sound exactly like you. Then it can have you "committing" genocides, molesting kids, or doing anything embarrassing, and send it to your friends and family.

At first this will be very traumatic for some people. Eventually, people may stop believing anything they see. Then nobody knows what's true anymore, because almost everything could be a lie. That makes it hard to know who to vote for, or if anyone committed a real crime.

It would also overwhelm law enforcement and courts: either you put everyone in prison all the time, or you let everyone commit whatever crimes they want, because nothing is provable. Then society starts looking like Somalia or Mexico, where organized crime is more powerful than institutions.

And unlike physical crime, you can't escape to somewhere else, because cybercrime is global.

This Isn't a "Terminator" Risk; It's a Human-Using-AI Risk

People are scared of sci-fi "Skynet" scenarios where AI magically comes alive and wants to kill everyone for no reason. But the more likely risk is evil human activities using AI to do this at massive scale.

Right now, cybercrime is mostly just slowing down economic growth. But even before AI, this parasite has been growing rapidly. Cybercrime already costs around $10 trillion per year—bigger than almost every country's entire economy. If nothing changes and current trends continue, the cost of cybercrime will exceed the entire global economy within several years. That means ongoing negative economic growth and everyone getting poorer.

Today, the Light Side is still more powerful than the Dark Side. But the Dark Side is rapidly draining the wealth of the Light Side. Once it has more resources than the Light Side, it becomes impossible to defeat, because it can fund more advanced AI agents that drain the productive economy—until there's nothing left to produce food, medication, or anything else.

We've seen this dynamic before in Somalia, where al-Shabaab controls large swaths of territory and rakes in millions of dollars from extortion, kidnapping, and drug trafficking.

Why Governments are Bad at Fighting Cybercrime

Government employees often get paid the same no matter how hard they try. And unlike the free market, government doesn't get the benefit of experimentation, where lots of people try different approaches, the ones that don't work go out of business, and the ones that do work grow.

It's like evolution: organisms get random genetic mutations without knowing in advance which will be best. The helpful mutations survive and spread; the unhelpful ones die out. That's why monopolies are bad. You lose that process of experimentation and survival of the fittest that channels resources toward the most effective organizations. Government is essentially one big monopoly, so it struggles for the same reason.

On top of that, governments are building AI weapons, but they can't even avoid getting hacked themselves. The FBI, the IRS, basically every agency is already getting hacked. A lot of their systems are old and outdated.

Why They Spend on Tanks Instead of Cybersecurity

Humans evolved millions of years ago, before computers, when we needed to fear violent physical threats from other humans, not cyber attackers harming economic growth. So people are driven by these instincts far more than by rational logic.

Ideally, governments would allocate resources based on the actual severity and likelihood of threats, the way insurance companies use math. But governments don't do this. They spend money based on how it makes voters feel.

It makes voters feel safer when a politician says "we're hollowing out our military if we don't build a bunch of new fighter jets and aircraft carriers." But people think it's boring when you talk about cybercrime.

Warren Buffett—one of the richest people in the world—has said that cybercrime and cyber-terrorism are the greatest risks to humanity, even greater than nuclear war. But the average politician probably isn't that smart, and they're the ones who get elected.

There's also an established military-industrial complex that's already set up to make aircraft carriers, jets, and bombs. It's very profitable, so they have tons of money to fund the campaigns of members of Congress. There isn't an equivalent set of cybersecurity companies with enough money to lobby that hard for security contracts.

The Solution: Make Ransom Illegal

This parasite needs resources. It gets resources through ransom. Every time a company pays $1M, that money can fund many more attacks.

If we make it illegal to pay ransom, then they have no money to commit more attacks, and less reason to attack. It becomes more profitable to do normal productive labor than to steal.

For non-ransom theft, we need much stronger penalties and a global system for identifying perpetrators and seizing their assets, so it's more expensive to do cybercrime than the benefits you get from it. Governments currently suck at this.

We also need a bounty system. The government should put bounties on cybercriminals and reward white hat hacking groups for identifying and neutralizing them, and for building AI to do this automatically. These groups are already doing it for the good of mankind; governments should be funding them.

If you pay based on results, you know the people you're paying are actually good at it. You give them enough to keep paying their people, and the better they get, the more they earn. That creates an evolutionary, competitive system for catching criminals.

This is basically what they did in the Wild West. Back then there wasn't a good police system or strong institutions, so you had bandits killing people and taking stuff, and the response was wanted posters with bounties.

My Personal Story

Let me tell you my tale so that you might avoid my fate. One of these parasites created an email address, mikepsinn@Outlook.com, and told me that if I didn't send him massive sums of money, he would send emails from that address pretending to be me and admitting to various crimes against humanity.

So if you ever get anything from me asking for money or claiming that I've committed any war crimes, please let me know at m@crowdsourcingcures.org.

Given that I barely have enough money to pay the $800-a-month rent on my one-bedroom apartment, I can't send him massive amounts of money. And since our government devotes almost no resources to solving these crimes, all I can really do is assure my contacts in advance that I have never eaten any babies or genocided anybody.

What's happened to me isn't yet incredibly widespread, because the "gestational period" of these parasites is currently 18 years, given that they're human. Unfortunately, we're soon entering an era where that gestational period will be close to 0 years, because it will be possible to infinitely replicate AI-trained versions of them, limited only by computer processors and electricity.

It will be far easier and cheaper to train AI on your public writing and social media posts, then have it generate believable "evidence" of crimes you never committed and spread it across social networks and to all your email contacts.

Genie vs. Golem vs. Djinn

Mankind has the capacity to be a genie, capable of things previously indistinguishable from magic. Something as simple as a pencil costs only $0.10, yet it required the intelligent, voluntary cooperation of thousands of people: to cut down the tree for the wood, to mine the ore for the chainsaw that felled it, to build the internal combustion engine in that chainsaw, to feed the inventor of that engine, to mine the graphite, to build the drill that mined it, to build the tanker ship that carried it across the ocean. None of these people ever met each other, and no central planner told them exactly what to do.

It's a magical, emergent phenomenon: humanity cooperating to become a metaphorical genie that has created wealth and knowledge unimaginable just a few lifetimes ago. It's a positive-sum game: you go to the store, hand someone money and say thank you, they hand you something and say thank you, and both of you are better off. The "double thank you."

The alternative is a negative-sum game played by golems. Factions of society form these golems, which just fight each other to steal power and resources while producing nothing but death and destruction. Nations at war each form their own golems, grotesque creatures fueled by hatred of the other golems, stomping on houses and crushing innocent bystanders. Since golems produce nothing of their own, they survive by sucking up the vast wealth created by the genie of human cooperation.

Yet golems aren't the worst entities. Their harms are usually the inadvertent results of infantile, short-sighted stupidity. The worst and most sinister entity is the djinn. The djinn is an emergent entity from the evil empire of parasitic criminals, with the explicit intention of draining humanity's genie of its power and starving it.

Where golems tend to just slow or stop progress, a massively powerful djinn will reverse it entirely, dragging us back to a primitive, pre-technological age as it drains hospitals, power plants, governments, and food and medicine manufacturers of all their resources, until every institution collapses.

The djinn plays a negative-sum game, the golems play a zero-sum game, and the genie plays a positive-sum game. You have to pick which of these three you'll be a part of.

AI Familiars: A Defense System

We need every citizen of Earth to have a global swarm of AI "familiars": little robots that inform each citizen of any verifiable violation of human rights and coordinate to impose consequences.

These familiars could:

Record what you do so you can disprove fake accusations
Keep data local/private (not on Google or Microsoft servers), accessible only to you
Coordinate democratically to gradually impose fines on malevolent actors
Help verify what media is real (trusted-hardware watermarks) so we can distinguish real evidence from AI-generated evidence
Run on trusted, unhackable hardware that can't be remotely modified

For example: in the case of Vladimir Putin's invasion of Ukraine, a global swarm of familiars could keep draining his financial resources until he can no longer fund the invasion and human rights violations.

This depends on there being more good actors than bad actors, and on people being able to access truth.

An Appeal and a Plan

I'd like to make an appeal to the cybercriminals of the world: come to the light side now, and avoid taking the world down this trajectory of an inverse singularity that completely undoes the massive human progress of the last two centuries.

I'd also like to appeal to governments to ensure that individuals who may have little legitimate opportunity for income are given real opportunities through positive job-creation programs that are more financially and personally rewarding than cybercrime.

We, the citizens of Earth, need to unite to defeat this shared enemy of every decent, productive person. This doesn't have to be a violent war. It has the luxury of being bloodless, yet it's still a noble and valiant movement. All we have to do is correct the incentives.

As Charlie Munger said: "Show me the incentives and I'll show you the outcome."

We need to make the cost of violating human rights greater than the rewards criminals get from violating them.

Decentralization is Critical

Centralized power is inherently limited, in both information and public support. And given human nature, if a controllable superintelligence is created and owned by some power-loving individual (the type most likely to seek power in the first place), we're in trouble.

Studies show a high concentration of psychopaths in politics; these people rise more easily than honest, decent, humble ones. So it's mathematically only a matter of time before someone like a Donald Trump, Kim Jong-il, or Vladimir Putin gains power and never wants to leave. If such a person controls a superintelligence that can conduct mass surveillance, destroy the reputations of political opponents, or kill them effortlessly, you'll have a permanent totalitarian dictatorship.

Concentrated, centrally planned power over AI is simply destined to destroy humanity.

You Should Not Be Hopeless

Global, decentralized power-sharing among all citizens of Earth is a good idea that would vastly improve everyone's lives. Communism was a terrible idea that vastly worsened everyone's lives, yet it still managed to consume half the world within a single human lifetime.

Given that, there's no rational reason a good idea couldn't be even more successful. The only thing that would prevent it is the irrational belief that it can't succeed. Purge that irrationality from your mind.

Even if only a small number of people can be persuaded at first, never doubt the ability of a small group of dedicated individuals to change the world. Indeed, it's the only thing that ever has.

Poll: Which Will You Be?

Do you want to become a magical genie, a moronic golem, or a parasitic djinn?

The Commandment of Wishism

Be excellent to each other.

Sources

Key figures cited above:

Further reading: