This vastly understates the work involved. Most C extensions and embeddings will require major structural changes or even rewrites. These things are everywhere and are a major reason for python's popularity. A typical financial instition, for example, will have a whole bunch of them, with a morass of python code built on top. Many of these companies took ages to transition to python 3 (and many still have pockets that haven't!). For them to remove GIL reliance from their C extensions is a much bigger ask than migrating to 3 was, and their response to being asked to do it is likely to be blunt.
The fact that only a small minority of developers involved in python directly use the C API (and fewer understand the consequences of GIL removal for them) means the issue tends to be overlooked in discussions like this, but it's the reason a GILectomy will be worse than 2 to 3. In practice it's likely to end up being either a fork or a mode you have to switch on/off, both of which would be miserable for everyone.
The GIL is part of python's success story. It made it easy to write extensions, and the extension ecosystem made the language popular. Every language doesn't have to converge to the same endpoint. Different tools are suited to different jobs. Let it be.