Hacker News Reader

I think about this constantly.

Not to sound like I walked uphill in both directions back in my day or something, but I remember building models in numpy without pandas. It was tedious -- and that's just a nice API wrapping ndarrays!

> Not to sound like I walked uphill in both directions back in my day

Local minima?

Most likely, gradient descent with momentum.

Oh boy, that and perturbation.

I’m not so sure.

You can make an argument current tools haven’t really surpassed a Lisp Machine for developer productivity, or a SmallTalk environment.

Not really. I see things like leftpad and npm fails and CEOs mailing private keys they've stored from customers. I see the same lessons we have have to re-learn year after year.