2. Telnet is, fundamentally, a remote login protocol; it's been bodged into doing other things, but that's its primary purpose, and it should be obvious why that needs encryption.

And you're seriously worried about wasting cycles on encryption? Did I accidentally step through a timewarp to 1970?

3. Telnet being available opens the attack vector of "user uses telnet to log into a remote system, transmitting credentials in the clear". That's a big enough vulnerability that the fact it doesn't introduce any other ones is really beside the point.

4. Telnet isn't just a program that can be used to do something unfortunate. It's a program that does something unfortunate by default, if you try to use it in the obvious way for the purpose for which it is designed. It's a gun that comes with the barrel pointing out the bottom of the handle.

And I really don't see any decent use cases for it. There are better tools for network testing, and better tools for remote login. It's time for it to die.