# Julia as a neural interpreter? Could this be a game-change for Julia and ML/AI?

**URL:** https://discourse.julialang.org/t/julia-as-a-neural-interpreter-could-this-be-a-game-change-for-julia-and-ml-ai/75386
**Category:** Machine Learning
**Created:** [January 28, 2022, 5:56pm UTC](https://discourse.julialang.org/t/julia-as-a-neural-interpreter-could-this-be-a-game-change-for-julia-and-ml-ai/75386 "2022-01-28T17:56:06Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![Palli](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/palli/32/3380_2.png) [@Palli](https://discourse.julialang.org/u/Palli)
#### Post date: [January 28, 2022, 5:56pm UTC](https://discourse.julialang.org/t/julia-as-a-neural-interpreter-could-this-be-a-game-change-for-julia-and-ml-ai/75386/1 "2022-01-28T17:56:07Z")

</div>

I’ve sometimes posted ML stuff I find intriguing to offtopic. This seems relevant enough to post here.

@ChrisRackauckas claims Julia doesn’t have an advantage for “standard” ML (unlike SciML). This may not be “standard” but if I understand if correctly, and this is the future of ML (replacing Transformers), or something in this direction, then I think the BLAS argument may be going away:

> **[Dynamic Inference with Neural Interpreters](https://arxiv.org/abs/2110.06399)**
>
> Modern neural network architectures can leverage large amounts of data to generalize well within the training distribution. However, they are less capable of systematic generalization to data drawn from unseen but related distributions, a feat that...

See here interview (I’m not yet there, I’m at this timepoint, that I think may be of interest to others):

[![](https://global.discourse-cdn.com/julialang/original/3X/1/1/11f013e289bf016b0027c1ebd636875a95e3498d.jpeg "Dynamic Inference with Neural Interpreters (w/ author interview)") ](https://www.youtube.com/watch?v=w3knicSHx5s&t=497)