[MathYouF]

I'd love a chance to talk to this guy (maybe at a rock-star after party at NeurIPS this year) because my view is:

1. "Either everything is magic or nothing is", and magical statistical tools are currently some of the coolest magic we have harnessed.

2. Making realistic pictures of "Corgi's with sushi" is cool.

3. Papers describing architectures which can make ever more realistic or interesting pictures of corgis and sushi are deserving of academic recognition, even if they can't precisely describe how, just as renderers which can do so would be.

I've found a lot of papers which primarily study the theory of ML coming out of the UK, A yet very few of them end up being of much value to the advancement of the field in terms of applications.

Conversely, the people focused on making real applications ("tools") seem to also be innovating on the science.

Tesla and Edison contributed massively to the progress of the study and application of electricity and both spent nearly no time in academia and instead focused on practical applications in industry. Edison's methods of investigation were said by Tesla (somewhat admiringly) to be entirely empirical. I wonder if the UK's strict academic approach to ML may be holding them back from making bigger contributions to the field. I'm glad some are trying to be strictly scientific about it, but I'm also glad it's not everyone, because it doesn't seem to be what's delivering the "magic" we all want.

[nth_order]

I agree with your views - its great and ultimately enriching that people approach this fascinating field on these varying levels of mathematical rigor. They complement each other nicely. He kind of says it himself, „we are building tools for people to get things done“, meaning there is this utilitarian point to ML, and empirism as a vehicle into the unknown is totally fine. The rigor/theoretical treatment will come over time anyways.

What I can more sympathize with is this notion of „AI colonialism“ happening somewhat. We can‘t trust our ML tools too easily, and should listen to domain experts maybe a bit more than many currently do.

That he criticizes AlphaFold2 is especially interesting I think. That model is clearly very useful, and crucially it is well calibrated and its limitations are in the process of being characterized. The measures of model quality are also not arbitrary at all, they have been used for sometimes decades in the field. If funding dries up - we will see. That would be tragic as there are still many interesting problems to solve there, AF2 should just be a start that enables research in many interesting new directions.