From Matt: This is a relatively short (3,000-ish words) piece for my premium subscribers. If you want to read it, you can sign up for $8 a month. Like I said in my last article, I plan to publish 3-4 premium articles a month, alongside my usual weekly free newsletters.
As always, if you want to get in touch, drop me an email at me@matthewhughes.co.uk or follow me on Bluesky.
A few weeks ago I was interviewed by a guy called Myles McDonough for a podcast he plans to launch next month.
I like Myles. He’s a Harvard-educated author that has done something I can only dream of — actually completed a work of fiction and published it. He shares my loathing of generative AI, and my dismay about the trajectory of the technology industry. He looks a bit like Oscar Wilde — if, instead of being committed to Reading Gaol, Wilde was the frontman of the Decemberists. And he asked some genuinely good questions that forced me to defend my views.
It wasn’t, as you might suspect, two “haters” talking about how much they hate something, never challenging the other.
Myles brought me to task on something I said in an earlier newsletter. Paraphrasing, he asked why I believed generative AI was harmful to the people who used it, and would it be harmful if it was… actually good?
I said yes. I believe that even the platonic ideal of a large language model — one that never hallucinates, and that wasn’t built on the wholesale expropriation of intellectual property, and that didn’t ruin the environment, and that doesn’t cost people their livelihoods, and brings about a golden era of productivity and abundance — would still be bad for people on an individual level.
You might wonder why “a golden era of productivity and abundance” would be so terrible. I’d say because such a thing would, naturally, have a cost — and I believe that cost is to strip away a key part of what makes our existence meaningful. It diminishes the very nature of what it means to be human.