`since it is trained to simulate the real world, as opposed to imitate the pixels.` It's not that its learning a model of the world instead of imitating pixels

`since it is trained to simulate the real world, as opposed to imitate the pixels.` It's not that its learning a model of the world instead of imitating pixels - the world model is just a necessary emergent phenomenon from the pixel imitation. It's still really impressive and very useful, but it's still 'pixel imitation'