I can't fathom willingly entering the web mines if you've already sunk in the hours to master high perf native code, OSdev, dist systems etc. People usually enter web to avoid having to learn such things.

To discard that knowledge and instead compete with an endless stream of fresh grads, boot camp grinders and LLMs for less money? A madness if I've seen one.

Web tech is more immediately rewarding. With C++ (and especially with modern web dev agile processes), you can spend a week building enough framework and infrastructure crap to stand up the barest inkling of a demo of the solution you plan to implement, only to have the whole thing pivot out from under you and become wasted work.

More complicated languages and frameworks that manage memory and do UI and everything for you let you get to a prototype much faster, so you can see whether it'll work or whether you want to toss out the idea much sooner.

It's not hard to find feature support or implementation differences between browsers. But these days, that's not really the problem. I can go read the ES language specification, and generally find user agents/node/deno/bun have compliant implementations.

The problem is brittle "ecosystem" stuff. How are the create-react-apps going? How is module bundling going amidst the UMD/CJS/ESM war?

If you want interesting, go to HPC large scale simulation.

Even if the decision he made was a bad one, you don't know anything about why it made sense for him at the time.

People aren't algorithms running in complete isolation and silence. They live in nuanced circumstances you can't possibly understand by his short history in a comment.

I hope you can find understanding and peace in the future.