Nowadays my two 'anchors' for handling feedback are the following:

1. Know what you yourself want to do and how you want to do it. Only you yourself can know that, and it is understandable that you have a vision about your project and how much changes you want/plan/can implement

2. That being said, the surroundings and 'how it goes' are quite important. You always should be open to feedback, and (at first) not even limit yourself too much in your own beliefs. There are always things you may have not considered, and you can always improve or take a chance that made itself available

To give some more context: I am currently planning out the projects I want to do with what I am doing here. Granted, it is only a hobby for me - so you already have much more space to maneuvre and do your own thing (so keep that in mind). The reason why I am doing that is because I want a clear idea and plan going forward, as it feels right to me to do so. With this statement alone you would already have a lot of situations/personalities who would strongly disagree with you, however.

Based on that, I also know what I want to put into the games, and how receiving feedback will be able to shape this vision. And while this does not mean that you could not potentially still go a complete open way of re-designing everything as it comes along to make it a better game, it is not something that I plan to do, simply believing that I will get out something enjoyable the way I already planned to do it.

And those last scentences above are where it becomes tricky, because as you said yourself: You believe to have talent - and you should (no matter if you could improve). So handling feedback and the way you incorporate can/will always get tested by certain situations where people would like to have you do it different for one reason or the other. Because they want to see it made another way, because they think they are right, because they believe you are not caring for feedback otherwise. Again, it is not that they have to be wrong - they could have valid points. It is just that thinking feedback has to be 'this' or 'that' is a way too limited approach. Some people/environments understand this, and they do not mind how you do your work as long as it is good. Sometimes you find yourself in situations where you are constantly questioned about your way of doing it - and it does not always have to mean that your work is not good. It can be quite a variying experience.

That is why I believe the first point to be important: Know yourself what you want and how to handle it. Because then you can ask a lot of people for feedback and take everything as valid. No matter what group of players, no matter what they believe and what they are providing. Everything is valuable, and everything that is not for you -personally- - it is your job to remind you how you would like to handle it. You do not have to change your gameplay if you do not want, even if you got 100 reports about it. But you can look into all of them with an open mind to see if there is not still something in it that you could see a point in - and maybe even do change in your project - as long as you want to do it.

The second point is more versatile and objective(?): Your surroundings play an important role. If you know you have people that play your games who understand you and your work and how you want to incorporate feedback, you can either let more into your testing, or simply write out to them what your are looking for (and what not for example - can save them a lot of work if they know you do not want to check that specifically). Also - workload. As you said yourself, it can be daunting to go through so much feedback. It is totally valid to reduce the amount of testers (or give them clear directions) because of that. It just comes back to the question: Are you limiting it because of a realistic circumstance or because you do not want to handle it as described above?

About your points specifically:

1. Feedback that makes your game universally better

While I see where your are coming from here (and you are basically correct), do keep in mind that even these things can be subjective. Not everyone is bothered by the same control schemes, mechanics (technical aspects), and some (perhaps you yourself) see them as solid and an improvement. So even this feedback needs to be split up, theoretically.

2. Feedback to make the game fit better with others of the genre

Yes and no, based on what I described above. I am personally more in the 'Yes' team myself, as I do not like to do the same things over again just because 'they are part of the genre' based on what people feel it has to contain. A game is a game, it can be enjoyable (for some), no matter how it fits into a glove. That being said, if you are on a budget and have to make something that appeals, this thought process can break your neck if you strain too niche.

That being said, remember the point about being open for yourself: Maybe they do mention things that you can see improving your game, or that they have a point of your game missing something - even if you want to give a different experience. Or maybe even reading their feedback or their point for a game mechanic gives you an idea how to change it up and include something more unique into your project that still qualifies why people suggested it. It can potentially be a win for both sides.

Other notes:

Technically speaking: Even a fps player could potentially give you valid feedback. For example about controls and camere movement in your first-person-puzzle-explorer. But yes, it does make sense do keep in mind where they are coming from saying 'the game has not enough action for me'. Still, even that could be something: Want to include something more active in your game after reading this? It may not be enough for the fps gaming crowd, but it could benefit your game how you envision it and give others more enjoyment. Again, stay open and think around the initial point of the feedback. But yes, if you only have so much time, focusing your testing on a more limited group - based on what you want to achieve out of that testing (there can be different 'studies' you want to perform), can be important, too.

About the positive things:

Debatable. Depending on how you define 'teaching'. You cannot specifically learn every talent, that may be true. But executing said talent efficiently (and even that word is debatable in what it could mean) is an ongoing learning experience. You can always improve over time - based on the lessons you have learned - and how you handled them. Of course, we once again have the two sides of the coin here: Yes, you can already have talent. Your work could already be good. And then good old fps gamer (no offences meant, they have their valid interest and place as everyone else) comes around and says your game is 'not good' because you do not take their feedback and have nothing to show to prove them different. You know what is different when you have made a name for yourself? Your experience? Debatably.

Yes, of course you got more experience. But what would be the first thing people may say about your work at that point? 'Ah yes, the developer has made a name for themself, they handle feedback like -this-, and people like their work.' Great. It just shows how empty these arguments can (not have to) be. You can already make something decent. So stick to it. But be open to feedback. That is how it should be, I believe. Everything else is just making it bigger in your head than it has to be.

Now again, if you do this as a job and got stuck into a scenario where your game is not receiving the popularity it needs to, it can get limiting. That is why you often read about specific design choices made because of thinking processes you as a player would not have thought to be the reason of why the game came out as it did. Or why it was only like 'this', or not like 'that'. And then sometimes a game became a masterpiece specifically because of those limits, and sometimes it flopped. And the other way around, as well. Which only brings us back to the two points in the beginning: Know yourself, know your surroundings. If you know that and why you have to handle things the way you do, you can - at best - handle it and yourself as honest as possible and combine making something you would want to do with how it may has to be based on how it is (or can be, if you are open to it.)

In short: I basically agree with you a lot, so my rambling is more of giving it a bit more 'insight' from my point of view.

Something I've learned as an artist is to take a critic's expertise and possible bias into consideration when receiving feedback. It sounds like a similiar approach ought to be taken when getting feedback from play testers, especially with things that are inherently subjective.