ML needs a new programming language – Interview with Chris Lattner

Weird that there has been no significant adoption of Mojo. It has been quite some time since it got released and everyone is still using PyTorch. Maybe the license issue is a much bigger deal than people realize.

I personally think they overshot themselves.

First of all some people really like Julia, regardless of how it gets discussed on HN, its commercial use has been steadily growing, and has GPGPU support.

On the other hand, regardless of the sore state of JIT compilers on CPU side for Python, at least MVidia and Intel are quite serious on Python DSLs for GPGPU programming on CUDA and One API, so one gets close enough to C++ performance while staying in Python.

So Mojo isn't that appealing in the end.

The problem I've seen is this: in order to get good performance, no matter what language you use, you need to understand the hardware and how to use the instructions you want to use. It's not enough to know that you want to use tensor cores or whatever, you also need to understand the myriad low level requirements they have.

Most people that know this kind of thing don't get much value out of using a high level language to do it, and it's a huge risk because if the language fails to generate something that you want, you're stuck until a compiler team fixes and ships a patch which could take weeks or months. Even extremely fast bug fixes are still extremely slow on the timescales people want to work on.

I've spent a lot of my career trying to make high level languages for performance work well, and I've basically decided that the sweet spot for me is C++ templates: I can get the compiler to generate a lot of good code concisely, and when it fails the escape hatch of just writing some architecture specific intrinsics is right there whenever it is needed.

The counterpoint to this is that having a language that has a graceful slide between python like flexibility and hand optimized assembly is really useful. The thing I like most about Julia is it is very easy to both write fast somewhat sloppy code (e.g. for exploring new algorithms), but then you can go through and tune it easily for maximal performance and get as fast as anything out there.