←back to thread

Speeding up PyTorch inference on Apple devices with AI-generated Metal kernels

(gimletlabs.ai)
186 points nserrino | 6 comments | 03 Sep 25 17:03 UTC | HN request time: 0s | source | bottom
1. earthnail ◴[03 Sep 25 19:11 UTC] No.45119420[source]▶
This is amazing. I wouldn't have thought that AI is this good in niche topics. Very impressive experiment and write up.

Still, I can't help but think we should bet on sth like Mojo instead for the long run.

replies(2): >>45119679 #>>45120295 #
2. ipsum2 ◴[03 Sep 25 19:42 UTC] No.45119679[source]▶
Mojo is a terrible language and its main feature (GPU acceleration through Mojo max) is closed source and requires a commercial license to be purchased.
replies(1): >>45120481 #
3. moelf ◴[03 Sep 25 20:52 UTC] No.45120295[source]▶
>I can't help but think we should bet on sth like Mojo instead for the long run.

JAX or Julia hopefully.

replies(1): >>45120496 #
4. earthnail ◴[03 Sep 25 21:15 UTC] No.45120481[source]▶
Why is it a terrible language? Genuinely curious question.
replies(1): >>45126207 #
5. ◴[03 Sep 25 21:16 UTC] No.45120496[source]▶
6. rovr138 ◴[04 Sep 25 11:47 UTC] No.45126207{3}[source]▶
A language where a main feature, Gpu acceleration, is being behind a commercial license is a pretty big issue.

Not on language design, but in the ecosystem and other areas.