I have doubts though about the alternative ways to experiment. Yes, I do believe that there may be a small, passionate community who would participate in that.
The majority of developers though (my assumption) have no interest in experimenting with features or sending feedback, they simply want to use a feature right now. So one discovers that -webkit-border-radius (old example) works quite well and without bugs, and from that point on, it stops being an experiment, it’s just a way to use something.
It’s in a way understandable that developers “pre-use” features like this: as you say, the web moves tediously slow and browser makers sometimes have odd priorities. Add in the competition from native apps, and developers feel frustrated and resort to experimental features en masse. It is at least perceived that we don’t have years to wait for features of which many can be considered basics. The world moves way faster than that and expectations are far greater, yet the web does not move at that pace, unfortunately.
That said, small-scale experimentation backed with solid test cases founded in the real world, may actually be a good idea, for as long as iterations are short and fast.
Finally, my hope would be that browser makers more closely align on priorities. It seems quite random right now.