Components

I was chatting to a friend at the weekend who works in a field about as far away from the hi-tech world of crypto as one could imagine. Her business, a traditional business with a heritage stretching back 100s of years, went “all in” to NFTs last year, believing it was the future of their organisation. They spent tens of millions of pounds training people like my friend (who has just about got her head around WhatsApp) in the murky world of non-fungible tokens, and building out infrastructure no doubt sold by people who promised the world.

This year the business is apparently on its knees. They spent money they didn’t have on an early-stage tech bubble. Non-fungible tokens may have a use in the future. They may even have a role to play in the future of art. But last year was little more than a proof of concept mixed with a pyramid scheme. It was not a technology on which to bet your centuries-old business.

I’m currently reading W Brian Arthur’s excellent The Nature of Technology, in which he argues that a particular technology is itself the amalgamation of a set of other technologies. NFTs are composed of a bunch of previous technologies like cryptographic techniques, computing power, the internet and so on. And those technologies themselves are an amalgamation of other technologies (computing power through devices all the way down to electricity supply, for example).

Reading this book has made me realise that the early hype waves of a new technology are usually a phase where the technologies in question are actually subcomponents awaiting a place where they might be useful. A solution not in themselves searching for a problem, but rather searching for a higher-level technology that itself solves a problem worth solving.

It’s a bit like inventing a hard disk drive and then creating a test rig that puts data onto the drive and takes it off again. It proves a point, but in of itself is useless.

Which, inevitably, brings me to large language models and generative “AI”. Chat GPT and all that.

These are very, very interesting technologies. But they are mere components. Toys searching for a place to call home.

There’s only one useful thing you can do with a toy. Play with it. But don’t bet your business on a toy unless you are Mattel.

The next phase of what’s currently happening is about understanding potential applications, not thinking that these new components are the solution for which you’ve been looking.