Taking VC money to being a better (probably worse) tmux is the sort of clownworld stuff only people in the Thiel circle can come up with
My take is that it will be tmux as a service. You will have many different sessions which will have their data stored in some central system (read: raft/etcd) and an API built on top of that for create/list/attach to sessions.
Ghostty is the client side aspect of this (the binary that actually runs on your laptop), but there will be a large backend system that the binary talks to through an API. With the API you can then also start automating sessions, for example when a metrics value trips an alert start an agent to triage with a session you can then attach to for further interrogation.
What Im not sure of is if they are going to build on top of kubernetes or not, which already provides a lot of the features necessary (etcd store, API server, workload orchestration, tenant isolation, extendable API).
(This is all speculation)
What does tmux alone do way better than the newer tools?
I have a lot of respect for Mitchellh but taking VC money for a tmux clone is wild.
As long as you are rich all the VCs will throw money at you.. fine.. but if you are rich then you don't have to take the VC money either. Just do it out of passion.