Describe the project you are working on:
Projects that require me to create many animations, tracks, keys.
Describe the problem or limitation you are having in your project:
Creating animations, tracks and key is a very tedious process. Especially so when trying to animate not just one property of an object, but a few.
Describe how this feature / enhancement will help you overcome this problem or limitation:
If we had a record feature for the AnimationPlayer, I could hit "record" even prior to any animations existing, make a change to any object, hit pause or stop when I'm done, and only fine tune the timing and values if needed. If I made changes to multiple properties, (like scale, position, rotation, texture ... what have you) all those tracks and keys would be created for me. If there was no animation in the AnimationPlayer, it would automatically create a new one and give it a dummy name.
Show a mock up screenshots/video or a flow diagram explaining how your proposal will work:

Describe implementation detail for your proposal (in code), if possible:
I only got a C-- in the "How to turn your computer on and off" class. So I'm not quite on the level of C++ yet.
If this enhancement will not be used often, can it be worked around with a few lines of script?:
I suppose it would be used a lot, as it would make animating anything a whole lot faster and easier.
Is there a reason why this should be core and not an add-on in the asset library?:
It's another UX improvement (a rather substantial one this time) to the AnimationPlayer. I don't think you should need go to the asset library to make existing tools efficient to use.
How does this differ from the autokeying option (the "car" icon at the top of the 2D view when AnimationPlayer is selected)?

Note that this feature is present in 3.1.1, but the icon wasn't added in time, which means you have to click the invisible space where the icon would be.
How does this differ from the autokeying option
I think it's about recording a set of values over time, i.e. you change property and it's added at current time, but time advances automatically (not sure if this is a case with current auto-keying)
@Calinou
Neither 3.1.1 nor 3.2 alpha2 Win64 has no visible or invisible car symbol up there, so I really don't know how to answer this question.

Even if such a feature already excited (which I cannot test, as it is not there), having it as (in)visible car up there is really of no good.
Please for the love of everything sacred, lets not question incredibly basic well known user behavior.
A red disc next to a play and stop is what everyone knows and immediately recognizes as a record button. Please, oh please, let's not make Godot usability worse by inventing our own little way of doing things when there already are well known world wide spread user behaviors conventions.
Found the car in 3.2. alpha2, not sure what I did to not make it appear before.
Anyway, so I can answer your question @Calinou:



So to summarize: That car does not look like a recording feature, it's not at the place where I would expect a recording feature, it does not perform at all like a recording feature, so in total it does nothing I would expect from a feature outlined in this proposal.
Found the car in 3.2. alpha2, not sure what I did to not make it appear before.
It was directly committed https://github.com/godotengine/godot/commit/296c74072c0f4d71567c425a94d9502a68b6e177 with the icon missing initially, then added https://github.com/godotengine/godot/commit/5d65100669e8a8ece09677feb744c69b4466aa90
Only the first commit was taken into 3.1,1 which went to release shortly after. So nothing you did or didn't do. It should be appearing in the next 3.1 release if the PR above is merged before then.
Also, someone once suggested this. Not sure if it went anywhere.

Most helpful comment
Found the car in 3.2. alpha2, not sure what I did to not make it appear before.
Anyway, so I can answer your question @Calinou:
So to summarize: That car does not look like a recording feature, it's not at the place where I would expect a recording feature, it does not perform at all like a recording feature, so in total it does nothing I would expect from a feature outlined in this proposal.