(In the past, people have dismissed such concerns, saying that it would be impractical for anyone to actually do such a thing. But this project by Intel shows that it isn't just possible--in fact, it's almost inevitable.)
The problem is GCC made these concerns completely irrelevant. You wanted refactoring support for C or C++? Either build it on clang or get screamed at by RMS for leaking GCC internals when text base search and replace was enough for him in the early 80s.
The simple fact is that there is no GNU based alternative to clang/llvm.
https://lwn.net/Articles/853230
Somewhere down the thread there are questions about AST, I think.
Sure, I’d love for Intel to upstream every optimization they make, but I don’t see that as necessary or even good business sense for either Intel or LLVM.
In any case, it sounds like you want to blame the llvm license for the bad feelings the student might have, but that blame is misdirected: the student should aim their anger (if any) at their professor or university for introducing unreasonable requirements. Or maybe they should suck it up because without llvm they would have had to pick another subject for their thesis. I suspect that llvm wouldn't have grown to where it is now (for better or for worse) with a copyleft license. And then our student would have been "forced" to contribute to gcc instead, and sign over their copyright to the fsf instead.
Finally I think the academic world is a lot like permissive licenses already. Anyone can read papers, use that information in any way, and there is no obligation to contribute back. Only when writing new papers are you expected to acknowledge your sources.
How are those PlayStation 5 optimisations working out?
Guess which compiler is currently lagging in ISO C++ 20 support.