About

Hi. I’m Irwin. I’m not a programmer but, as a programmer friend of mine said once, I know enough to be dangerous. I want to share and document my experience trying to wrap my head around what Large Language Models can and can’t do. I have a background in design, interaction and UX, writing, investigation, and journalism (among other things) so I’m specifically interested in the following questions:

• Can we train AI to competently do things that designers can do?
• Are AI agents really possible? Can they really do useful things for me?
• Can they do legitimate, trustworthy investigative work?
• What ethical and societal implications are there, if the answer to any of the above questions is “yes”?

To be clear, I approach these questions as a skeptic — an open-minded and curious one — but a skeptic nonetheless. I want to see for myself and I welcome you to follow along yourself if you aren’t already and share any discoveries you make along the way.

I hope you find some of this worthwhile, enough to spend some of your spare attention on it. I’m not going to open up comments from the outset, but if you want to post a reaction you can do so on Saturation or Bluesky. This is mostly for me, a way to impose some of the discipline of thinking and writing on myself. It’s not to cultivate an audience or sell ads or make a living.

That said, if you would like to buy me a taco or support independent social media, the Fediverse, and the original ethos of the WWW, you can contribute to the administration and running of Saturation.social through my Patreon and if you would like to join Saturation, it’s a lovely bunch of people, and I’d be happy to create an account for you there.