3 comments

[ 110 ms ] story [ 731 ms ] thread
This was discussed earlier in the year: https://www.modular.com/mojo

The fact that it's apparently closed source / proprietary, plus stuff like the obviously bullshit 35k times faster claim make it a tough sell imo.

Personally I see ML migrating to lower level frameworks rather than more abstraction.

Hey if it contains some ideas to make actual Python faster then I’m all for it.