In my last essay, We’ve Always Been Taking Off, I wrote about how humanity has been on an exponential path for a long time. The singularity is just the next step in that curve.

People define a singularity as the point where intelligence becomes self-improving. The thing is, that’s already happening.

Coding is the clearest example. Developers aren’t working alone anymore. We’re writing code with systems that learn from our code. Every prompt, every bug fix, every repo makes the next model better. If you work at an AI lab or do ML, you’ve probably solved a matmul issue with AI. It has probably also caught a bug for you. That is self-improvement in the loop, and we’re in it.

We’re not observers waiting for the singularity to happen. We’re part of it.

I’ve seen it in my own work. My learning speed has gone up several times since using AI daily. Not just for technical work, but for everything that helps me be better at it: leading projects, running experiments, collaborating across teams. I’ve learned a lot of that through ChatGPT. Those gains feed back into how well I can contribute to building better AI. It compounds in both directions.

That’s the real shift. Humans are improving AI, and AI is improving humans. The loop is already closed.

So the singularity isn’t a future event. We’re already in it. The real question now isn’t when it happens, it’s why. What does it mean to be human in a world where intelligence is shared? What is the goal of all this progress?

Answering that might be the key to understanding what true alignment actually means.