←back to thread

ML needs a new programming language – Interview with Chris Lattner

(signalsandthreads.com)
311 points melodyogonna | 1 comments | 05 Sep 25 11:33 UTC | HN request time: 0s | source
Show context
postflopclarity ◴[05 Sep 25 13:59 UTC] No.45138679[source]▶
Julia could be a great language for ML. It needs more mindshare and developer attention though
replies(4): >>45138887 #>>45138911 #>>45139421 #>>45140214 #
numbers_guy ◴[05 Sep 25 14:18 UTC] No.45138911[source]▶
What makes Julia "great" for ML?
replies(3): >>45139000 #>>45139316 #>>45142213 #
postflopclarity ◴[05 Sep 25 14:26 UTC] No.45139000[source]▶
I would use the term "potentially great" rather than plain "great"

but all the normal marketing words: in my opinion it is fast, expressive, and has particularly good APIs for array manipulation

replies(1): >>45139131 #
numbers_guy ◴[05 Sep 25 14:38 UTC] No.45139131[source]▶
Interesting. I am experimenting with different ML ecosystems and wasn't really considering Julia at all but I put it on the list now.
replies(2): >>45139169 #>>45139451 #
postflopclarity ◴[05 Sep 25 14:42 UTC] No.45139169[source]▶
Glad to hear. I've found it's a very welcoming community.

I'll warn you that Julia's ML ecosystem has the most competitive advantage on "weird" types of ML, involving lots of custom gradients and kernels, integration with other pieces of a simulation or diffeq, etc.

if you just want to throw some tensors around and train a MLP, you'll certainly end up finding more rough edges than you might in PyTorch

replies(1): >>45148775 #
1. salty_biscuits ◴[06 Sep 25 12:40 UTC] No.45148775[source]▶
Yes, my experience has been that it is great if you need to do something particularly weird, but less smooth to do something ordinary.