It ossifies every layer of the stack below it. Once you have a practical monoculture, user demand effectively cements the entire underlying operating system design to cater to the browser implementation. Fine if you're selling Chromebooks, but fatal for innovation.

Once the operating system is encased in amber, only a massive budget can break new ground -- this is why things like Plan 9 from Bell Labs will never take off. It's why Haiku is importing more and more BSD code instead of pursuing their own paradigms. It's why the only realistic chance for anything new is Google's Fuchsia; they're one of a very few organizations who can afford to innovate in this environment.

> It's why Haiku is importing more and more BSD code instead of pursuing their own paradigms.

We are? Uhh ... not really?

We use FreeBSD's WiFi stack and drivers (same as the other BSDs), and have plans to import some of FreeBSD's work to get linuxdrm/gallium for hardware acceleration. That's it.

Probably a good thing on the surface - less work for devs in terms of compatibility checks. Add competition has made all browsers less complacent, improved performance everywhere, sped up adoption of new tech and general been an adrenaline shot. Much less talk nowadays of “will browsers support it?”