This guy has over 46 thousand citations.
He has extensive expertise in DL especially in areas around vision and object detection both of which are going to be crucial as part of building World Models. Look at his work on MobileNet if you doubt this.
He can also run large teams of brilliant engineers, despite Apple failing to launch a product we know there were issues internally that didn't land on the feet of this person.
The salary is obscene, but the potential upside for Meta of getting to World Models first is nearly unlimited. Zuckerberg is making the right move here.
5 minutes of research gets you this level of information, although some domain experience is required to understand the ramifications of these things.
It's easy to bitch in a public forum ignorantly but it's incredible seeing how wildly wrong people are lately, fairly understandable given the topic but I think there is also something else going on: a general feel of being replaced and obviated.
Separate Apple's perceived AI failures (how could Apple as we know it EVER launch an LLM – they won't even allow explicit content on the App Store!) from the rapidly evolving industry and don't keep your head in the sand or you'll be buried.
This technology is not a panacea and has drawbacks, limitations, and flaws, but it is not comparable to anything in recent memory.
Most of us including me won't like where things are going, but that doesn't change reality.