You'll probably consider this another "power play", but if you find yourself agreeing with Tyler Rogoway then it's a good indicator that you need to re-examine your position. Rogoway isn't as bad as Kyle Mizokami or Minnie Chan, but that's a very low bar to clear.
Substantively, the premise that the J-20 was designed to "shoot-and-scoot" is contradicted most strongly by the J-20's characteristics themselves. Why would it include control surfaces like the canards and ventral strakes for improved aerodynamics if all it needs to do is fly fast in a straight line and shoot? Especially since such surfaces negatively impact RCS to at least some degree, and a plane defenseless in WVR needs all the stealth it can get.
This was touched off by the claim that the J-20 can fly really fast. Great! It's an air-superiority fighter that can fly really fast because of its wing sweep and manoeuvre really well because of its use of canards, LERX, strakes, and an advanced lifting body. That speaks to the brilliance of Dr. Song - he was able to find a region of the design space where a highly stealthy fighter with outstanding speed and manoeuvrability could be conceived.
I would argue that the climb at 4:25 is pretty show-stopping, not because the angle or speed are particularly aggressive, but because of the remarkable coupling between the canards and main wing through the LERX that the condensation demonstrates:
Even more impressive is that this was done with a mediocre engine without TVC. When the WS-15 is ready, the J-20 will be the closest thing there is to a perfect 5th generation fighter.