Paging @Leguark

Suggestions to improve the article

Update code snippets for latest (significantly different) GemPy 2.x version
Update the two plots with better figures.
- Figure 1 doesn't really serve much information - I suggest it can be cut down to one clear view of the data in 3-D. Maybe add a 2-D slice plot instead of the 3 different camera views.
- Figure 2 should contain model fault displayed
- clean marching cubes "artifacts" (cut surfaces with fault)
- White plot background for printing?
In the second to last paragraph you switch from geomodeling to "geomodeling as an inference problem" without much explanation. I doubt most readers will know why automatically differentiable software is key for that and what that all even means. I'd recommend first just talking about stochastic simulations in general, and how running GemPy on GPUs can speed things up. Maybe then hint at the ML use cases afterwards. If you're running into the word limit, maybe cut all the references to the widely known open-source packages used in GemPy, which are probably out of scope for such a short article.

I'd be happy to make those changes if you'd like 😄

softwareunderground / 52things