Large Language Models, aka LLMs, aka AI, aka ChatGPT, are bad for society. I don't want to beat around the bush on that, but I also don't want to talk about things like the destruction of art, the cratering of code security or accuracy, or the generation of large emails from bullet points that are reduced to bullet points on the other end, causing a lot to be written while data is lost. I don't want to talk about how we generate billions of photographs every day, billions of articles, and billions of songs, too many for any human to experience, so many there's people thinking about a data storage crisis, but now we have computers generating more without a reason. No, what I want to talk about is at the heart of organizations and people.
See, I've come to see LLM usage in organizations as the opposite of signing a union card. When a worker signs a union card, they are making a strong statement. They are saying they are workers who love their job. They are saying they love their community and their livelyhood, and that they will fight to be allowed to work to make their community better by doing the work. A union card is a signed statement that a worker wants to make the world better and will fight for that. Using an LLM is the opposite. When a worker uses an LLM to do their job, be it coding, or writing an email, or especially a creative work, they are saying they don't want to do their job. They don't want other people to do their job. They want their bosses to fire them and destroy their community.
You might disagree with me on the value of unions themselves, or the value of LLMs, but I think that's a red herring. Because I'm talking about the actual act of joining one. People join unions because they want to ensure a better future. It doesn't matter if you personally think that this is the correct way to ensure a better future for them, this is something they are doing. One joins a union because they want to keep doing the work they are doing, people rarely start a union for a field they hate working in, they'd rather just leave the field (where possible), they start a union to keep working in the field, but better. And due to the threat that most executives/bosses feel from unions, signing the card means you are going to fight for your job. It's a sign you want to build a better work community.
On the other hand, if you are a programmer and you decide to get Copilot or ChatGPT to write a function for you, then you are making a statement that the LLM should replace you. It's not even that the LLM should replace you - it's that the LLM should replace those like you. The others who do your job. It's telling your boss that you and your colleagues don't matter. If you use a model to build an image for your blog header - then you are saying that it's fine to fire artists, that the work the artist did was valuable (you want that header), but the artist themselves don't matter. That the people don't matter. You are asking for their community to lose a person who is support. Turning a person who is contributing money and effort to the community, into a person who needs help from the community. There's a stat I hear a lot - that every job created in a community is actually many, because several people can work supporting the people paid by the outside jobs. But we need to think about the opposite, replacing a person with an LLM means destroying multiple jobs, means destroying the restaurants, theatres, book stores, and more that make up the place you live.
I think that LLMs are bad, I do not think that editting with ChatGPT, or getting it to create boilerplate like test suites for you is effective. I think that using an LLM to do work will make you worse off in the long run. I think that using an LLM to learn is the simplest way to hurt yourself as a person. But I don't think that's the worst part. I think it's a statement that learning from someone else is something you don't value. It's a statement that you hate your own community. And I think that we, as a society, need to teach this to others, learn this, and experience this, and share this.
Because we can't just secure computers