Dr. Whitfield Diffie on The Future of Cryptography and the Threat Posed by Quantum Computing

Pure luck. I had a long-standing interest in cryptography that was ignited in 1972 by a difference of views between NSA and ARPA.

I left a steady job at the Stanford AI Lab and began traveling around, digging up rare manuscripts, talking to anyone willing to talk, and thinking about the requirements for a cryptographic system.

That journey led me first to my wife of 43 years and then to Martin Hellman at Stanford. Marty and I ended up working together for four years and what came out of that was public key cryptography.

I did. When she was interviewed by the RSA conference, my late wife remembered my telling her that by the end of the century, people would be having intimate relations with people they had never met and that cryptography was the only way to secure their communications. She said she didn’t believe me, but she ran off with me anyway.

The change in scale has been incredible. Back when the military was the biggest client for cryptography, a large order was maybe 100,000 machines. Today, high-grade cryptography runs in billions of browsers.

At MIT, I was working on symbolic mathematical manipulation — teaching machines to do calculus. I didn’t consider it AI then, but in hindsight, it was.

What grabbed me at that time was the problem of program reliability. I imagined mathematical proofs of program correctness rather than endless debugging.

That may still be the most important problem in engineering, but it hasn’t made nearly as much progress as cryptography, and I was lucky to have been drawn away from it.

Several things. First was David Kahn’s book The Codebreakers in the late ’60s which sparked interest in the field. Then IBM began working on cryptography for banking applications and brought in Horst Feistel, who had a developed block ciphers (what we use today) for the U.S. Air Force in the 1950s.

That work led to the Data Encryption Standard (DES), which made cryptography available to everyday business. Finally, our work — public key encryption — offered a completely new mathematical framework, which attracted mathematicians and opened the field in new directions.

It appears real, even if the timelines are uncertain.

Experts differ — Adi Shamir thinks it may take until the end of the century, others say we’ll see it by 2030. But the business point is that NSA and NIST take it seriously.

They’ve developed new post-quantum cryptographic standards, and there’s now a 10-year timeline for government systems to comply.

That creates a clear roadmap — and a business opportunity — for everyone who needs to remain compatible.

It means change — very significant change. The new cryptographic standards are computationally different. They cost more in terms of time and memory, and nobody’s exactly sure how the transition will go.

Major cloud providers like AWS are already working on implementation, but we might see alternative approaches emerge mid-transition, just the way public key cryptography did back in the Seventies. So follow the standards — but stay alert for new ideas.

Biotech will change the world even more than computing. On one hand, biotech will change our very selves. On another, gene-synthesis labs are getting cheaper and more widespread. Controlling who can access them will be impossible.

I see a window of vulnerability where it will be easier to create new diseases than to develop treatments. That may be an existential threat to civilization.

AI, meanwhile, may accelerate everything.

I used to believe humans would no longer be controlling the world by the end of this century. Now I think it might be by 2050.

Stay flexible. Work to implement the new standards but be open to alternatives. Post-quantum cryptography is going to be a dominant concern for the next decade. It’s not clear that the current solutions are the final answer.

And don’t forget: the rest of the information security stack — implementation, operating systems, communications protocols — is riddled with bugs. Crypto is the best-cooked part of information security, and everything else needs even more work.

Because we’re entering a new phase in the story of digital trust. Cryptography is evolving again — under pressure from quantum computing, driven by AI, and intertwined with the growing complexity of our digital and biological systems.

If you’re responsible for securing information — or building services that depend on it — this matters to you. I’ll be talking about those trends, the direction we’re headed, and what it means for everyone in this business.