https://openssf.org/press-release/2021/10/13/open-source-sec...
Edit: if you're talking about obtaining certs for bare IPs, it's actually possible to do that. See 1.1.1.1
My guess is that OP's point was that "everyone" includes servers as well as browsers, presumably by shaming via no lock icon and other tactics.
For most of us that means an untrusted cert, and hence a browser warning, but that's equivalent to the browser padlock icon when browsing http (and arguably safer).
Plus, if you are browsing or publishing using plain ip addresses you are already outside the "regular behaviour" so users doing that will be happy to click through.
However this subthread is irrelevant to the main thread. The https/http question is unrelated to the dns question, since dns applies to both equally.
This will not work anymore as the https will require to have an official/qualified domain name.
Look at the case of the .dev domain names that were used by a lot of persons, and suddenly google bought the top domain and enforced https mandatory for it in chrome to ensure that no one else could use it anymore.
Also, imagine if the next step is to check that only "approved/legitimate" top domains, respecting some Google policies, are allowed for your own safety...
1. This affects http, too. I mean it affects anything using domains names.
2. It's generally a bad argument as there are so many more ways state agencies can mess with you when you use http then when you use https.
3. did you read the article? (I guess no)
4. The problem is DNS and DNS only, everything else is just a side effect from DNS being conceptually broken (wrt. privacy protection and decentralization). Https everywhere is inevitable and good, what isn't is that due to how DNS works we ended up with a (kinda) centralized-only certificate system. Https happens to need certificates, but it doesn't care what system provides them for it. So a complete privacy preserving decentralized certificate (and DNS replacement) could totally work with HTTPS.