If you look at the cutting edge of browser tech lately, and you read between the lines as to why the browser makers are doing this, you can see that there's a pretty significant trend now towards browsers actually fighting the web itself, because once multiprocess is done and modulo a 2 or 3x rendering speedup that Servo may be able to bring (which will probably be the last advance of that size), the only way for a browser to feel "less klunky" or "speedier" than its competition will be for it to render less of the web. That is, the classic optimization of doing less work.

Which in this case means trashing all the shitty ads and plugins from 15 different sites and synchronous JS trackers, since that happens to be what's klunking up the web. Which, fortunately for the browsers, gives them cover to do this under the guise of working for the user, because there's a lot of truth there, but in some sense it's still just cover, too.

It is in some sense a change of an era; browsers are moving out of a phase of enabling and empowering content creators and slowly moving into active conflict with them. As slowly as possible in some sense, because it's scary for them, but it's something they're being forced to do. It's the same effect where computer users will blame "Microsoft" or "Dell" for the computer being slow, when the problem is actually a slow hard drive or something; eventually it stops mattering that it isn't "your" fault and you have to do something about it.

Big changes in the browser space in the next five years here, I think. Content producers so far have only tepidly fired back at ad blockers and stuff because they're still very niche, but they are getting less niche very fast now.

> once multiprocess is done and modulo a 2 or 3x rendering speedup that Servo may be able to bring (which will probably be the last advance of that size)

That's what we said for years about computing power and the Moor law's : "transistor can't get tynier than xxx nm". It happened anyway because there used to be that much economic incensitive to improve computer. Now research in micro processor is so expensive that the gain in computationnal power are found elsewhere. But there is always ways to improve things. The question is : "is there people eager to fund research". As long as the answer is yes, we'll find ways to improve web browser.

In the web there is so much money involved that the answer will still be yes long after servo shipped.

There is no ability to eternally speed up the browers without doing less work (in the sense in my comment), and while it may never technically limit to zero, it will plateau at some point. Just as Javascript performance pretty much has.