On: The beauty of functional languages in deep learnin...

[jackpirate]

As the author of HLearn, I 100% agree. I do all my actual ML work in python (which is most of my work these days).

At this point, the Haskell ecosystem just doesn't have the engineering support of something like tensorflow, and so there's no competition.

[dmix → jackpirate]

Community size always outranks the pros/cons of languages. Still they are never entirely always static and it always takes some bold people to push it towards something new and (hopefully) better. Although more often than not it stays the same.