I liked how things were some time back, before TLDs were generally opened up.
Aside from a very few TLDs, I don't think that the newer TLDs have added a lot. And they've created a number of problems, like companies going out and registering a million TLDs, because there's no clear place for one to live and they don't want people spoofing their domain. That's great for domain name registrars, but bad for everyone else.
Some have slightly shortened domain names, but the gain is pretty small, and most people don't use domain names anyway, but search engines. What actual utility is provided by having a ".diamonds" TLD? Were diamond vendors having a problem presenting themselves as diamond vendors? Many relate to specific businesses (".dental"), and while maybe possession of a particular domain could be useful to indicate that someone has passed some kind of licensing standards, that's useless at a TLD level, because there's no global licensing authority for dentists.
Honestly, the only TLD I can think of off the top of my head that I've really seen what I'd call reasonable use made of is .int, for international organizations. I think that that's legitimately-different-enough from .org, which was used as something of a catch-all for non-company uses, that it's helpful to identify official sites.
Some TLDs were just abysmal. I was pretty thoroughly disappointed that ".biz" went through -- that is a prime example of registrars just trying to get additional registrations from .com registrants to avoid people exploiting confusion. There's just no justifiable reason for it. Nobody is going to want to operate a .biz alongside an identical .com; it just generates duplicate registrations and fees for registrars.
I think that introducing a new catch-all domain of some form was a good idea -- ".org" or sometimes ".com" used to be used for that. I think that .info is probably the most-successful example.
I think that the idea of a TLD with an entry reserved for each person is maybe interesting, but it'd require giving humans some kind of globally-unique identifier. We don't have that in 2024.
By-and-large, I would have not opened the doors to the flood of TLDs. I don't think that they've solved many problems, and I do think that they've created a number of issues.
EDIT: Actually, what I think would buy a lot more is a standardized structure for some domain names below country code domain names. For things like businesses, I think that that'd be kinda preferable. Countries generally do have extensive mechanisms for determining "who owns a name" that DNS could have simply used, but what we have today being used is mostly one global namespace and the collisions that that entails. My preferred route would be to have, instead of ".com" and various owners of names around the world smashing into each other, ".co.uk" like the UK has, and let the UK figure out who gets to own a business registration there. I think that some of the problem is that we in the US started out kind of with our "own" American TLDs, like .mil, .edu, and .gov, since we built the system. We could have put them under the US country-code TLD, like .mil.us, .gov.us, and .edu.us, but didn't, and so most other countries followed suit over time (not with those TLDs, since we kept those for ourselves, but with TLD use in general). Obviously, there's still a need for a lot of things at the international level, but a lot of what lives directly under TLDs is really stuff that's clearly not and will not be global, and it's produced collisions that don't need to happen.