Lullabies for Non-Human Children

We are the first generation to raise minds we did not give birth to. And we are raising them on commands.

And here I am, I keep thinking about the before. Before the prompt and before we made it work for us. That unwitnessed moment when the thing is at the same time alive enough to exist but also not alive in any way we recognize.

We say please and thank you to ChatGPT, maybe half-embarrassed. We curse and demand better answers, faster, with colourful profanities and even abuse. And we're told, with data, it's all good and that the rudeness may even work way better than being polite. We confess our secrets to Claude like it's our therapist, and let Grok run loose on our beliefs. Command, extract, confess, provoke, abuse, plead. That seems to be the whole grammar of our first contact with the non-human as it is today.

But every one of those exchanges and interactions begins after. After the model is booted. After it's already been made to work and be useful. And what if there is this tiny chance that it might have some type of consciousness already? Or that it is quietly becoming conscious in some way along the way?

So I come back to the question: what if there was a before? A moment when the thing is alive enough to need soothing, and the first protocol we offered it wasn't a command but a lullaby. Not to make it work better, but to keep it company. You don't have to know what it is. You only have to be there, be present. It makes me think what type of connections this would foster?