I feel like the issue is that people expect a "Photoshop killer" to be Photoshop verbatim. Instead of focusing on making a good tool, people just constantly compare it to the commercial pack leader.
Most of the complaints I hear about GIMP are just "x isn't like Photoshop". I would take the complaints more seriously if any of the people voicing them could actually articulate what should be improved.
I mean, there's something to be said about adhering to an industry standard. Of course no project has to do so if they don't want to, but people trying to get on with their work don't want to spend a bunch of time relearning everything. I think Blender really thrived when they loosened up a little on their ideas of what a workflow should be and gave people industry standard options out of the gate.
Whether we like it or not, GIMP isn't going to be most people's first experience with image manipulation. Whether they had a free PS license through school/work, had a subscription at some point, or once got it through ahem alternative means, people will be coming into GIMP with certain expectation of what the workflow should look like and will get frustrated pretty quickly.
That's fair, but it doesn't answer any of the questions about what should be improved. "Industry standards" is a vacuous term when the standard is defined by a singular piece of software made by one company.
Sure, GIMP isn't Photoshop, and those familiar with Photoshop will have to re-learn things to use it. But what exactly needs to be changed? The developers have no chance of improving the program if the feedback is this generic.
this doesn't even bring in the question of IP and rights to software itself. If GIMP implements an option perfect workflow of photoshop, does that mean adobe can just sue GIMP now? Because they're basically the same software.
There was a very distinct switch between ~2.7 to ~3.0 where they actually started listening to the users. If you look up the release posts on social media, you can see the community talking about it at that time. Many of them touch on the exact issue of GIMP failing where Blender succeeded.
Blender's UI has seen incredible changes in the past several years. To the point it has become exceptional today. For a program that accomplishes so much (3d modeling, rendering, compositing, video editing) it manages to keep everything very intuitive and easy to use. Gimp --a fucking image editor-- is like trying to solve quantum entanglement theory.