Please can you elaborate on what do you mean with 'too clean'? If you do so we may be able to explain to you if it can be done or no in real time or if it's something that can only be achieved with CG.
Some parts of the video seem to be cutscenes running on real time in the console (the traditional ones that in the past were using specific assets or lighting for cutscenes), and other ones seem to be that 'light cutscenes' using the gameplay assets (like the one showing the mamooth or when Aloy is riding far from the camera).
I think they will be able to achive the stuff we seen in the game running cutscenes in real time. Maybe not at native 4K or not at 60fps, but I think it's doable. Specially by guys like Guerilla, who like Naughty Dog or Sony Santa Monica are always pushing the limits of how good things can look in games.