|
Quick, what are the prime factors of 226,679?
Give up? All right, I’ll tell you.
The answers are 419 and 541, but don’t worry if you struggled: factorising large numbers is something humans find difficult. When very large numbers are involved—ones that are hundreds of digits long—computers find it difficult, too. That difficulty is what guarantees the security of most modern cryptographic systems, which protect online communications, financial transactions and corporate secrets. To crack them open, a hacker would need a new kind of machine that was fiendishly good at factorisation.
Unfortunately, such machines are already being built. Quantum computers, which exploit the strangeness of the subatomic realm to crunch numbers in powerful new ways, find such problems very easy. Because quantum computers do not really work yet, your secrets are safe for now. But if and when they do start working, perhaps in the next decade or two, the world’s cryptography systems will suddenly be worthless, and any secrets currently protected by them will be laid bare. So a new kind of cryptography is urgently needed that relies on more difficult problems than factorising large numbers. This is known as “post-quantum cryptography” (PQC).
Good news came last week when NIST, America’s standards agency, announced that three algorithms had been approved as official standards for PQC.
I wrote about
how the algorithms were chosen, what nearly went wrong along the way and what happens next. It turns out that some digital systems have already been upgraded to PQC, including iMessage, Apple’s messaging platform for iPhones, iPads and Macs. So if you use one of those products, you are, without realising it, already living in the future, which is pretty cool.
Upgrading everything to PQC is going to take many years, though. But we need to get on with it, because if you talk to people who work in quantum computing, as I do quite often (The Economist hosts an annual conference on the subject, which I chair), you’ll get the sense that the pace of progress in the field has picked up in the past year or so. Quantum computers still don’t work well enough to do anything useful, but when they do, I suspect things will start moving extremely fast. That was, after all, what happened with artificial intelligence (AI).
Like quantum computing today, AI was something people spent years working on without much to show for it. And then, a bit more than a decade ago, things abruptly changed, and the modern AI boom began. I wrote about this sudden shift, and what caused it,
in the first of our weekly schools briefs on AI,
which we’ve been publishing over the summer. The sixth and final chapter, on AI regulation,
was published this week.
So if you want a concise, accessible primer on AI, how it works, and where it’s going,
all six chapters are now available.
And who knows? Perhaps, in a few years, we’ll be publishing a series of schools briefs on quantum computing, how it is changing the world—and how we switched over to PQC in the nick of time.
Elsewhere in The Economist:
-
Listen to
two
podcast
episodes that explore a looming crisis in cosmology
-
And
read about
the intense cyber struggle between Iran and Israel
Thanks to everyone who wrote in about last week’s newsletter on AI and the brain. Addy Pross, a chemistry professor, sent a thoughtful note supporting the position of Francis Crick, a molecular biologist, who was critical of the comparisons made between artificial and biological intelligence. “AI is algorithmic (no awareness), while BI is cognitive (aware). AI has no goals or desires; BI does. AI is not alive; BI is,” Mr Pross said. “Francis Crick was spot on yet again with his comment that neural networks were biologically ‘unrealistic in almost every respect’.” If you have more questions about quantum computers, AI or anything else to do with this newsletter, reach us at
sciencenewsletter@economist.com.
|
