Kai T. Chen

I didn't pay much attention to CS or DS until I got here. I really enjoy the vibe and the student support at Berkeley, and I feel very grateful for almost everything I've run into so far.

Somewhere along the way I noticed my taste had shifted: I get more excited by things people made than by truths that hold perfectly. Gödel is where that started. I used to believe there were truths we could reach and hold on to, and the Incompleteness Theorems were where I first set that belief down — and where I first wondered whether being incomplete is really a flaw at all.

Which is, I think, why CS caught me. It's human work through and through: built, historical, contingent. And the longer I look at it the more I find to love — designs of real beauty like the Curry–Howard correspondence, small ideas that do far more than their size suggests (like chmod), and the blunt practical moves people make because something broke once and they remembered. That used to look like a downgrade to me. Now it looks like the interesting part.

And then the AI revolution showed up and went straight for that new footing. Machines can make things too, and they are very good at hitting whatever we choose to measure — so most of the evaluation systems we built for ourselves are turning into Goodhart's law right in front of us. Most AI safety talk I hear is about whether these systems might end us. I've come to suspect the quieter risk runs deeper: a world where everything we once measured ourselves by can be satisfied without us, and we slowly forget what the measuring was ever for. If the measures we used to judge human work by stop measuring anything, what was it about human making that I found worth loving in the first place? That's the question I keep coming back to: GOD holds. What else is there for us to seek, and to find joy in?