Today is one of those days when people might *not* look at me like I have three heads when I say disabling root certificates from smaller CAs in countries you don't do business with can be a good mitigation against MITM by shrinking the pool of roots an attacker would have to compromise to make a cert you trust.
Like a lot of countries have private CAs for signing their own government websites. How many of you would have problems if your users couldn't reach the Turkish or South African government? If you're not in one of those countries, probably not a lot.
The whole PKI notion, to me kinda always rubbed me the wrong way? Trust, wait, to whom is trust given to implicitly that I do know?!
(Open)SSH's or SILC's TOFU (Trust On First Use) paradigms, seem to avoid a lot of those sorts of pitfalls; though I suppose it creates a little more friction for users? TOFU paradigms also: don't act as a source of profit as the TLS/SSL CA infrastructure did (until Peter Eckersley [RIP] et al created Let's Encrypt to attempt to mitigate that anti-pattern).
It was kinda "hilarious" when some colleagues of mine managed to get a CA to issue a certificate to *. (e.g. https://eprint.iacr.org/2009/111) what could possibly go wrong?! 😉
Short Chosen-Prefix Collisions for MD5 and the Creation of a Rogue CA Certificate