The main issue I see is ease of MITM for corporate environments. In a corporate environment a trusted root is installed, then an appliance can intercept all SSL certs and re-create the trust chain to introduce their own trusted root so they can read all SSL traffic and your browser says "SECURE". That is broken IMO.
It is very difficult, if not impossible, to protect against attackers who are in a position where they are able to install their root cert to your browser. If they can do that, then they can do a lot more also. At least in the current system such tampering is generally easily detectable.
It's broken that device owners can make them behave according to their intentions?
Shall we eliminate software freedom too? Otherwise companies can just install a fork of Firefox/Chrome with MITM support added back in to their managed endpoints.
Are there any conceivable architectures where whoever owns the computer couldn't MITM themselves? If only for debugging purposes, which would immediately be used by corporate IT for the usual purpose.
