At least making what Oracle is saying mod-able sounds very very reasonable. (and making the amount of joints mod-able as well while we're at it)
If only, forearm 90º rotation and thigh rotation would be a great addition for broader movements with only 8 more joint states to be added.
"Competitively", it's jumping from 20 to 22 joints. I don't see how that'd change the needed reaction time from 60 seconds or so, to 5 minutes.
As for the weird DM's that could come out of it, we're going to have a whole new engine.
DM's can be finally be newly coded into something more predictable and without the weird little glitches that we currently have.
The increased difficulty for new players is something that actually matters, though.
That in particularly is what I truly hope they focus on for this new "game", with a lot of new approaches that "hold the hand" of people who open the game for the first time.
If a lot of strategies towards that are employed, in my opinion, it isn't the addition of 2 new "joints" that will make the game more unplayable than it already is.
Back when there was a "FRAME" line limit for every replay there was a lot of people that said you didn't need to make replays with more than what already existed.
Same goes for the old limit for objects in a mod.
Working those out increased the complexity of the game by a lot, and allowed multiple MB replays to exist as well as amazing mods that I'm sure would have made a huge difference back when we actually had a bigger userbase.
Right now the only way to make it happen is by using an extremely tedious hack that is just not feasible without writing scripts specifically for it.
I don't completely advocate for the addition of those 2 joints to the default Tori because of the "new-players" issue, but there is a lot of potential behind allowing the number of joints to be mod-able and allow what Oracle described to be possible.
tl;dr: Make the number of joints(/bodyparts) mod-able in order to allow what Oracle described to be possible.