Well, I would say it depends on the use case? Maybe for accessing a couple services I would be okay w dynamic dns, but if I’m trying to host something where I’d prefer less points of failure and prioritize stability and uptime I probably would rather have a static address. (Please correct me if I’m wrong)
Static IPv6 isn't that much harder to memorize. Admittedly I'm often numerically dyslexic, but I even sometimes have a harder time with memorizing IPv4 addresses than IPv6. Particularly with the big trick that you can elide all the zeroes you want with double colon (::). At that point you just have to remember your subnet prefix and then all of your statically addressed devices you can think of as just {subnet}::N. Your web server can be thought of only as "::1" to you and your database server as "::2" and your combination printer and coffee maker can be "::3" to you, and you just have to remember to include your prefix up front (or copy and paste that part). Even a paltry /64 is just 16 hex numbers to memorize in nice 4 groups of 4 in the worst case that your ISP can't give you a bunch of trailing zeroes in your /64 request. (A /56 which is what most ISPs should give you per convention if you request a static subnet is 14 hex numbers in 3 and a half groups of 4.)
The point of having a static IPv6 address would be that you can use regular DNS, not dynamic DNS. With dynamic DNS you have to use a short TTL, increasing the load on the DNS server, and even then it will break occasionally due to caching when the IP address changes.
Ease of memorization is a non-issue since you should be using DNS anyway, whether for IPv4 or IPv6.
The IPv6 prefix I get from my ISP is almost entirely "full", so the :: is only good for the latter 64 bits. I can imagine someone paying to get more zero bits in the prefix, or a more memorable prefix.
For me I've just resigned to the fact that each of my IPv6 servers will require DynDNS.
