AI vs Numerical PDE - Turbulent Euler/Navier-Stokes - Data Bank

Today the mathematics community has been struck by an OpenAI repository with mathematical proofs of mathematical theorems:

• Several hundred claimed resolutions of long-standing conjectures across number theory, algebraic geometry, representation theory and mathematical physics...
• Milne's rationality conjecture, Goldfeld's conjecture, Nagata's conjecture, Bloch's conjecture for complex surfaces, Fujita's freeness conjecture, Hilbert's tenth problem over ℚ, the irrationality of Catalan's constant. Those are not routine results; several have been open for decades.
• But nothing whatever in numerical analysis, finite elements, error control or computational PDE. That is not a subject being served unevenly. It is a subject that does not appear.

Here is a reflection on this new revolutionary state of affairs: 

• AI as Computational Mathematician Rather Than Theorem Prover

It seems AI as LLM can do analytical mathematics since it is a symbolic language, and then better than human mathematicians, like computer chess vs Karpov. But computational mathematics is a collection of well defined algorithms not asking for AI as LLM, and so can serve mainly for coding as a language, data collection and output evaluation. 

This means that computational mathematics appears as the winner under the AI advance, while the effect  on analytical mathematics may be profound. 

I will now test if AI for Euler/Navier-Stokes can deliver a data bank of representative turbulent flows with dual-based output error control, by running the same code with data collected by AI, and report the result.