Computer Club!

16 Sep, 2024

The research centre I run, the Centre for Culture and Technology, has its first session of a new program this evening. Computer Club is this new program, which I will be running in collaboration with Matt Nish-Lapidus. Computer Club is an outgrowth of two things:

For the past two years, we've run a program called Media Arts & Studies Playtime, which was a hands-on, low-stress, informal demo series, featuring artists and academics with something practical to share. We ordered pizza. It was fun.
Apple announced its Apple Intelligence. Microsoft, Zoom, Google, Amazon, Meta, and others had their AI Advertising Olympics. Wildly, Mozilla has decided to try to turn itself into an AI company. AI is eating the tech industry.

Others do really excellent work critiquing the tech industry and its AI dreams. Read Timnit Gebru, Alex Hanna, Ed Zitron, Paris Marx, and Cory Doctorow. (And I do have an essay in the works that I hope to submit in the next calendar year--but that's about the history of the critique of AI, which is much older than you might expect.)

I want to figure out how to divest, as a user, from AI. My first post on this new blog is about that desire.

That's what Computer Club is trying to do: work together with develop some fantasies, games, intellectual habits, and practical skills that allow ordinary(ish) computer users/nerds to compute without AI. Because apparently every browser vendor and maker of mainstream productivity software is all in on AI, that will mean, in 2024, that computing without AI means something like computing without the tech industry.

There are a lot of possible ways to go about doing this. Tonight's session will be about asking participants what they want to do! But of course, as the convener of the thing, I have to have some ideas. I want to write them down (in public); I also want to solicit other ideas. (Ping me at @scr@assemblag.es on Mastodon.)

In no particular order, here are some things we might do:

Play with an analog computer. (We have one at the Centre.)
Work through nand2tetris.
Make some generative art without AI, instead using algorithms instead of GANs.
Learn how GANs work by building some, e.g., by working through The Little Learner.
Develop a library of software and howtos dedicated to FLOSS alternatives to commercial software (e.g., Firefox [or, I suppose, LibreWolf] instead of Chrome; LibreOffice instead of Microsoft Office or Google Docs; Thunderbird instead of Outlook; etc.) This might focus on academic work.
Explore alternative internet protocols, e.g., the Fediverse, Matrix, Gemini, and so on.
Teach and learn how to use the command line and plain-text computing, including things like writing in markdown and using pandoc to generate other file formats.
Highlight the differences between small, simple tools you're in control of (say, text editor + markdown + pandoc [what I use] vs. the LibreOffice word processor).
Play with alternative computing stacks, including uxn, LISP machines (or their contemporary approximation with Guix, Emacs, and Stumpwm), various retrocomputers and emulators.

There are also three projects that have nothing to do with AI, or divesting from it, that I still hope we can do, at least one of them, at least a little bit:

I have a 2D plotter on the way. We will set it up and play with it.
I would love for us to build a polar plotter.
I want to build a Logo robot turtle.

These three projects are related to my history of computing work, studying Logo and Papert at MIT, which has informed my thinking here, deeply. But more about that later.

I have the distinct impression that my (aesthetic and/or political) commitments around computing are going to entail more than just using different software. To the extent that open source software is the alternative, there are two competing tendencies:

Community. Open source software is often community-based (although there is, of course, corporate open source--just ask IBM and Microsoft). Using open source software, whatever the model, involves doing something different and more than simply consuming software, even if it's just placing a feature request or reporting a bug on GitHub.
Autonomy. Open source software means you get to do what you want with the code. That does, however, mean that you sometimes have to do much more with your computer than consuming software.

Over just the past several days, I've read things that suggest pretty starkly that Mozilla has decided to turn itself into an AI company, adding the foundations for a built-in AI chatbot to Firefox. This is terrible; Firefox is what I use to be able to divest myself of Apple, Google, Microsoft. Of course, the thing to do, with enough open source moxie and knowhow, is to start working on a non-corporate browser, one that is free from whatever corporate brainworms are steering the ship towards the AIceberg. But your web browser is likely the most complex piece of software you use on a daily basis, excepting, perhaps, your operating system.

My hunch, and I'll have more to say about this later, is that actually computing without AI means computing without the tech industry, and that actually means computing with software small enough to recreate without the resources of an enormous corporation. Alternatives should be plausibly buildable. (The Ladybird project is laudable; for a long time nobody thought it could be done. They're predicting an alpha in 2026.) Hence the focus in the list above on small-scale computing. The scale of the web birthed the datasets large enough to train image generators and LLMs; web search and the web browser are now host to most of the consumer-facing AI in the world. Perhaps divesting from AI means divesting from the web. I don't know.

I do know that spending time finding out is the work, right now.