It seems to me that nowadays Webkit/Blink has replaced IE for developers not caring about interoperability – with some even going so far as to claim „IE? but nobody uses that!“.
All browsers do this. Firefox implements around 88 properties (30 of these are "logical properties", which aren't exactly new properties) that Chrome doesn't[1] and Chrome implements around 68 properties that Firefox doesn't[2].
The numbers are a bit skewed here. My method of collecting the list of "properties that Firefox implements" included preffed-off properties, but the same is not true for Chrome, so for example the `grid` property shows up as implemented in Firefox but not Chrome even though both have it behind a pref/flag.
However, there are some properties implemented in Firefox without being behind a pref (ruby-align, scroll-snap-foo, image-orientation are examples).
-------------------
I do feel that in general Blink does push a lot of nonstandard (or not on the standards path) features into the web at times. But I don't have anything to back that assertion up.
Blink is being interoperable in this particular case.
[1]: https://manishearth.github.io/css-properties-list/?stylo=sho...
[2]: https://manishearth.github.io/css-properties-list/?stylo=sho...
The problem arises when the experimentation is done with entirely new features that the browser wants to push (features which haven't been discussed and somewhat settled on with other browsers). If they aren't behind a flag, sites start using it, and then you have to support the legacy syntax forever.
clip-path is on the standards track. It's not going to change much. So browser vendors exposing it by default is fine.
(However, if you're building an experiment using these, at least test in other browsers and put up a warning that it won't work in some browsers. A lot of the interior issues today are because folks only bother to test in chrome. Ultimately this means that all browsers have to copy Chrome bugs to make things work.)