You aren't being attentive enough for high risk activities. You lack proper verification channels, to confirm authenticity, which matters in this context. You lack the situational awareness to proceed safely.
Admit that you might not be cut out for what it takes to maintain a secure posture on the internet, if that's what gets you. Just stop pretending to try.
And I never claimed I was good at anything. But I did intend to point out that with advanced persistent threats in the mix, and nation state actors operating with effectively unlimited resources, failing to notice the difference between two vaguely similar nonces (in a situation where it matters) is going to get you hanged, depending on what you're trying to fly below radar.
onion.nytimes.com CNAME nytimes3xbfgragh.onion
Edit to clarify - more a technical pondering than a solution to anything
I guess in theory a browser _could_ support something like that, but it'd be pretty unusual. I also think the idea of relying on DNS to resolve a hidden service would defeat a lot of the privacy and security guarantees associated with those services, so I don't think any browser serious about security would implement something like that.
Websites redirect all the time, there's 3 HTTP status codes for it.
If you visit https://onion.nytimes.com/ and it sends you a 301 redirect to https://nytimes3xbfgragh.onion/ then yes, I'm pretty sure that'd work fine. However, if you perform a DNS lookup on `onion.nytimes.com` and receive in response a CNAME record pointing to `nytimes3xbfgragh.onion`, I seriously doubt the browser is going to respond to that by establishing a new Tor circuit to the named hidden service. Rather, it's most likely just going to do what every other DNS client does when it receives a CNAME record; it'll try to look up `nytimes3xbfgragh.onion` in the DNS. (And fail, because `.onion` is not a valid TLD in the regular DNS system.)
I just generated "omen coins car hoof.onion" right now, in a couple of seconds, with my laptop CPU. With a couple of GTX1080s you could easily find some much better .onions than the ones NYT chose.
Lets pretend we've got a $5000 budget.
Quick back-of-the-envelope math shows that a 8xGTX1080¹ box will be able to generate ~3 onion addresses beginning with "nytimes" every second, we can afford 6.25 months of this.
Instead of waiting a really long time, we'll rent multiple boxes and squeeze all that into one month. In that month our servers will find approximately 44055283.1 onion addresses beginning with "nytimes".
At least with the wordlist² I use, without a prefix I discover approximately one "good" onion per 8 million random onions. Considering we've got a 7 character prefix for our 44 million onion candidates, so I'd expect a significantly better rate than just 1/8000000
So yeah, in a month you'd probably find at least 5 "better" onion addresses.
¹ $800 a month https://selectel.com/solutions/gpu/ Shouldn't be a problem for NYT to get a few of these.
² which besides words contains some easily memorable fillers such as "aaa" "bbb" and so on