While i do think the threat model of web-pki TLS is different from TOR, with sone of the problems in 1 not present in the other and vice versa, I'm generally unconvinced it is the pancea that this article presents it as.

> DNS hijacking is impossible. DNS is simply not used.

Sure, but now you have a new problem that tor hidden service identifiers are not really human identifiers. This makes attacks where you trick the user into going to the wrong site much easier.

Which is more likely - DNS hijacking + no TLS (or at least no HSTS) or a user being tricked into typing the wrong incomprehensible string of letters. Personally i find the latter to be the more realistic threat to the average user.

> BGP hijacking is impossible. Every interaction a Tor client has with a relay or onion service is authenticated such that you are guaranteed to be interacting with the relay/onion that you intend to be.

I dont understand. BGP attacks in order to do passive monitoring seem just as do-able for TOR, although i guess now you need to do it in two places.

Isn't bgp hijacking (if you can do it arbitrarily, which is unrealistic) basically a global passive adversary - the main thing tor famously doesn't defend against.

> There are zero places a corporate firewall can inject itself to decrypt the traffic. There are zero places and zero parties between the Tor clients that a MITM attack can be performed.

They could be the bridge. They could just spy on the client directly (usually the assumption is the corporate firewall can install stuff on your computer like custom CA certs. If push comes to shove just install spyware directly)

The main benefit tor has against corporate firewalls is it is to obscure for anyone to care.

> Do you assume the CT lookup process is secure? Are the parties you're communicating with misbehaving? Are your lookups in CT logs being logged and associated with you?

Admittedly im not super familiar with CT validation, but i was under the impression that certificates contained the SCT, which is a signature that is validated on the client and that there is no (online) CT lookup . So i dont understand - how can the CT lookup process be insecure or privacy violating if it essentially doesn't exist. All that is happening on the browser is a signature validation - no additional network requests needed.