What Computers Can and Can’t Do Efficiently

Why this matters more than quantum computers right now

Many people in crypto are talking about “quantum-safe” or “post-quantum” signatures because quantum computers could one day break some current cryptography. But:

-Useful quantum computers for breaking crypto are still years away.

-They will be specialized, expensive machines in controlled environments — not something anyone can just rent or buy like a normal computer.

-Governments and regulators will have ways to limit access, at least early on.

AI is different. It is already here, widely accessible, improving quickly, and hard to control. The ETH result shows that AI can suddenly change fundamental assumptions in math and computer science. That can shift the ground under cryptography and security without warning.

The risk of rushing into new “quantum-safe” schemes

Newer, more complex algorithms (especially ones that haven’t been studied for as long as classic ones) are more likely to have hidden flaws. AI or new mathematical insights can reveal those flaws suddenly.

A real example already happened: HAWK, a quantum-safe signature candidate going through the NIST standardization process, was withdrawn after AI helped find a serious weakness in it.

Critique of what many crypto chains are doing

A lot of blockchain projects are adding support for quantum-safe signatures and heavily promoting it. The argument here is that this is mostly marketing, not careful engineering:

-Adding a new signature type is relatively easy work.

-It lets projects show photos of quantum computers next to their logos and look “advanced” and sci-fi.

-Many token holders and traders don’t deeply understand the technical issues, so the narrative sells well.

-Unlike past safety scares (e.g., Y2K) that mainly cost money, poorly chosen new crypto schemes can actually put user funds and security at risk.

In short: chains are selling “protection from quantum computers” while the more immediate and harder-to-control threat (AI-driven discoveries) is already active. This prioritizes narrative and token interest over careful risk management.

One important distinction

Some serious organizations are working on post-quantum encryption for a different reason: the “harvest now, decrypt later” threat. Powerful adversaries (e.g., nation-states) might store encrypted messages today and decrypt them years from now once they have suitable quantum computers. That is a real, different risk for long-term secrets. It is not the same as the marketing-driven push for quantum-safe signatures on public blockchains.

DFINITY’s position

DFINITY will focus on the real, present risk from advancing AI rather than joining the quantum-safe mania. They intend to take a careful, responsible approach instead of chasing the narrative.