I hate to bring bad news on a Monday, but… 

Your money sits on a hunch.

Every time you log into your bank, buy a stock or send an email, a math problem stands guard. 

The bet is simple: certain cryptographic problems are too hard to solve. 

Factor this number. Find that point on a curve. Untangle this lattice. Solve one and you hold the passwords, keys, and access codes. Fail and the vault stays shut.

The entire global economy is built on a hunch that everyone—or everything—that tries will fail. 

Here’s what almost nobody knows. We never proved those problems are unsolvable. 

We assumed it. Cryptographers and hackers—maybe millions of them worldwide—spent decades trying to break these schemes. 

They failed. So we declared them safe. The digital economy rests on the collective shrug of a few thousand very smart people.

That worked. Until now.

No, It’s Not Quantum (Yet) 

AI has started doing real mathematics. Machines are now verifying proofs of famous results that most humans can’t read. 

One reportedly found a counterexample to the Jacobian conjecture, a problem mathematicians assumed would always hold. 

If you zoom out, the arithmetic is brutal. 

Our confidence in cryptography equals the hours humans spent attacking it. A single month of AI compute could match a century of that effort. Nobody knows what falls out.

Mainstream media has barely written a word. They don’t know where to start.

Now the good news.

It Hasn’t Happened (Yet) 

For years, experts predicted AI would unleash a hacking apocalypse. Planes grounded. Banks frozen. Despite models getting extremely good, it still hasn’t come. 

Some experts have offered a convincing reason why: Human hackers are diverse. Each one thinks differently. Meanwhile, AI models think alike. They train on similar data and reach for similar ideas. 

If North Korea and Russia each spend a million dollars on AI attacks, they’ll mostly find the same holes. Teams of AI + in-house experts spending two million finds them first.

Translation? In an AI arms race, money and expertise talks. Defenders usually have more of it.

Which brings us to crypto…

Crypto’s Strange Advantage

For your bank, security is a cost center. For crypto, the math is the product. 

Break the signature scheme and the coins belong to whoever broke it. So crypto has faced this threat first, in public, with real money on the line.

Watch what it’s done with it. 

Some projects built swappable cryptography and post-quantum signatures years ago. Many have been paying hackers and AI to attack their own foundations before anyone else can.

Sure, institutions are arming up too. Banks are testing AI that hunts their own weaknesses. One industry program surfaced more than 10,000 serious flaws since April. Google’s AI agents helped Chrome ship over a thousand security fixes in a month. One bug had been hidden for 13 years.

But look where they aim. Wall Street points AI at its code. Crypto points it at the math under the code. A bug gets patched. A broken assumption takes down everyone who leaned on it.

Finding the hole is half the job. The other half is closing it fast.

Who Wins? 

Systems built to resist change are strong in peacetime and brittle in a fast war. Blockchains that bragged about never changing now need to move. So do banks. So do governments. 

So who wins?

The owners of the defense. Compute. Verification tools. Security budgets. Companies and protocols that can upgrade fast. 

Bet against the ones that can’t.

Now more than ever it’s important to differentiate between the two. Because the big hunch held for fifty years. AI is about to check it. 

Odds are the math survives. Either way, the people who prepared will own what comes next.

More on that to come.