So if they have solved the problem why are we still complaining about it? All you need to do is show a big warning message for domains that are younger than 1-2 days.
If there were some way of bootstrapping new domains with some sort of positive trust signal (e.g. links from domains that have been in use for a long time?) then it might be possible to preemptively blacklist domains, as you say, but there are potential anti-trust issues with any such system that dissuades people from accessing the websites of new companies, for example.
This could be sold as an add-on for a certificate, or something like that, with a just high enough barrier for proving authenticity. Known-good domains could then additional treatment, like a blue padlock or one with a star (ok, I'm not an icon designer).
Of course you can argue against it on a freedom basis, but I think for protecting vulnerable web users it'd be pretty useful.
There's other tricks that scammers pull though, like compromising legitimate domains with a certain reputation and then adding their scammy subdomains, but that's a lot harder to pull off than just registering 100 free domains everyday and hoping one sticks.
The problem is always the political one. What does, say, Microsoft vouching for a site actually mean?
This is a classical arms-race and will only intensify. With domains that look generic enough and only serve malicious traffic when hit with the right URL, parameters, user-agent and geographical location, blocking will have to rely on sourcing these URLs directly from the targeted endpoints (e.g. SMS/WhatsApp/Email), rather than "crawling" or relying on users to report these. Another approach is to do some of the blocking locally, which of course means pushing the detection logic to the client and thus exposing the classification mechanism. Neither approach is sustainable long-term in my opinion.