If you are doing 3d graphics in Rust/wasm32/Chrome, is there any advantage of using the WebGL2 api over the wgpu (targettting webgl2) api instead ? Having played with both, wgpu seems strictly "safer/more structured" than webgl2.
Wasm still has an interface overhead talking to JS to talk to webgl, but it's not too bad when there's no strings or objects identities. We've long been promised interfaces in wasm which would close that hole, and wgpu can also target WebGPU which is now available in chromiums, so yeah, not very much other than the usual "it's more setup", which is only true for very simple JS apps.
I will say, if it's the interface specifically, wgpu is directly based on the WebGPU standard, so maybe you could argue directly using that in JS is better?