AI defeatism

A month or so ago, I spent four days surrounded by nice, caring people who combine the desire to do cool, creative things with computers and the passionate belief that the way to do that involves a lot of thought not just about software and hardware, but also about values and politics. I only heard three people use the term “vibe code,” only saw two demos containing some kind of “AI,” and only witnessed one (polite but firm) discussion about why a prominent software project absolutely does not and will not implement code written by large language models (LLMs). Maybe the less inspiring conversations were happening elsewhere and I wasn’t hearing them. The conversations I was present for gave me hope that the world I love isn’t gone yet.

Last week, I attended an event that had absolutely nothing to do with technology. I got to chatting with someone who was utterly convinced that we have to go all-in on “AI” because the horse has already left the barn. A new and powerful technology exists. Bad actors will use it. Because bad actors will use it, everyone needs to use it. This, you might guess, is not logic I buy into. To be honest, it’s also an argument that gets made in relation to just about anything that can be used with malice. It’s very much the “good guys with guns” sort of argument – if guns exist, good people should be allowed to legally own and carry them so they can stop bad guys with guns. And well, we know how that one goes: laws that allow good guys to have guns generally cause there to be more guns in society. More people are then hurt or killed by guns. Laws which strictly limit the availability of firearms don’t totally prevent the presence of guns, but they do tend to do a pretty fine job of limiting the number of guns on the street, and making it harder for "good guys" to shoot people. The answer to a technology which can be used for ill is not getting a lot more of it and putting it in the hands of people we think are somehow more moral.

When I scroll through my Mastodon (where you can find me, if you haven’t already) feed, I see that an increasing stream of people who have spent years doing cool things with computers are feeling a sense of despair. Computers aren’t fun anymore. The pleasure of coding is eroding. Basically, everything is getting worse. The despair is often accompanied by a great sense of sadness. There was a good and interesting thing that has been ruined. It has been ruined by a way of working that a lot of people seem to think is just going to happen, and keep happening, whether we like it or not. It is a distressing mood, and feels profoundly disempowering.

I don’t know many people who are genuinely enthused about what’s happening right now. Sure, there are people who use tools like Gemini or Claude or Copilot, and who feel that they have uses, or that they are somehow necessary in order to be able to be part of the future. But I don’t personally know anyone who loves LLMs and their ilk, and genuinely finds them nice and desirable to use. Like probably everyone, I’ve of course read/seen/listened to the stories about people investing emotions in chatbots. But that’s not the kind of love I mean. The kind of love I mean is the type that I personally only encounter in re-posted posts from work influencers in my LinkedIn feed: the people who are so enthused by the latest agentic whatever that they want to shout from the rooftops about how much better it makes their lives. I do not know any of these people personally, but apparently they exist.

While I do recognize that being engaged in communities where people develop software out of a sense of joy, love, and purpose (or just a desire to scratch their own itch) makes me both an edge case and very biased, I find it desperately sad that “spicy autocomplete” (as the cool kids call it) is taking the pleasure out of things people used to enjoy doing. People who make software because they enjoy it or find value in the process are being confronted with both tools, and systems of organization, which force a very different view of the act of developing something.

I would rather be like the principled software developers from the first paragraph, who know how they want to do things, feel strongly about what is and is not good for their project, and mobilize their arguments clearly and firmly. I would rather be like them than like the gentleman at the non-tech event, who has given up already and feels that we simply must go along with what is happening. To have already given up, to accept that this is our world now, feels like a supreme triumph of marketing and hype over humanity and democracy.

Instead of giving in to despair, giving in to "AI" marketing and accepting that this is life now, I want to believe that dissent is still possible and fruitful. Literature is full of nice ways of describing the contrarian, the fighty, or the hopeful-against-the-odds. I would rather tilt at windmills and be considered a fool; be the little boy who presumably got yelled at for pointing at the naked emperor; not go gentle into anything, never mind an allegedly good night. I want to continue to fight against exploitative business models, environmental degradation, the devaluing of labour. I want to fight the feeling of despair and sad inevitability. In short, I’d rather tilt at windmills now, with the possibility of being wrong later, than give up already, now, when there are still fights to fight, opinions to change, and despair to counteract.