llms

Now, I know what you may think: "Oh look, another blog post about AI and LLMs... how boring." I know, I know. Just open the Bear Blog discovery feed and take your pick. But first off, this is a Makko site, therefore: bnnuy (by one of the makers of Makko!). And second, this a take I hadn't really seen before and it might be my favorite take on the topic.

So please, go read it.

my philosophy on banning AI in projects

– a post by @alexia

What I love is that, in this post, she makes a call for a more people centric internet and codebase management. One where connection between people is valued above all else. That the fact that the knowledge gets shared from one person, to another, is special and should be treated as a precious thing. And how even the most of ethical of large language model would still get in the way of that.

So many projects and so much of the world would be better with this approach in mind, both now, and before the rise of LLMs. In fact, I'd say that it's because this approach wasn't the most common one that we ended up with this rise all together.

So much of the open source world, notably, has fostered a community that is downright hostile to pretty much anybody, that people would rather robots to do the work than contribute themselves. Robots don't care when they're being abused. They don't tire. They don't burn out. They don't cry.

Newcomers get ridiculed and shunned for their mistakes, constantly raising the bar of entry and therefore, the weight put on the active contributors who get burnt out until they themselves snap, and get bullied out of the codebase. The cycle continues, on and on until who can be surprised when all we have left are robots to do the work because we couldn't be bothered to show some understanding toward each other?

We value productivity over passion. In this sense, LLMs were inevitable. They're the logical conclusion of this apathy.

The promise of LLMs is to "make you better", to "enhance you", it's about "fixing yourself" to "compete," to better "fit in" and "integrate" yourself. Everybody has to be amazing at everything. You can be a one coder army with an LLM. You can do everything, on your own, you see? Within projects, this plays on our fear to ask for help and gives us a workaround, instead of trying to heal the source said fear. It's hyper-individualism disguised as self-reliance.

But it doesn't have to be this way. Things can be different. It's a choice we have to make. It always was. You don't know something? One of us can teach you. You can't do this? Someone else can. Knowledge is meant to be shared, disabilities are to be accommodated.

We need more communities and spaces in which people feel comfortable speaking up and asking what they would call "dumb questions".

...did I tell you I really liked this blog post?

#FuckAI #AI #LLMs