Not even a little, it's wishful thinking. No city is going to want a bunch of autonomous vehicles without proof that they won't be a danger to its people and traffic system. Imagine what will happen when Teslas start hitting people or start to stall traffic because they can't handle the situation. And that's just two variables of many that still need to be solved. Also, Musk likes to make predictions on time but he has yet to meet any of his predictions in terms of time. He reminds me of a boss I once had that liked to predict how much time a project would take yet he had no concept of the true time. He would promise a project for a certain time and we would look like jerks because we could not meet it. Eventually, we all got fired because the IT department was ineffective.

It might happen but not in 2 years. My guess is that we won't start to see auto autonomy until 2024. It's my guess but I bet mine is better than his.

I doubt a million but someone will allow it. I think they easily have the best solution and its going to take years before car manufactures put sensors suites in cars that would allow them to gather data to even get the navigate on autopilot level.

Waymo with lidar might have a chance, but the more I see the more people who've tried lidar says the pure computational is much better than the hybrid approach. Lidar only gives depth and it very error-prone. It's easy though because it requires no data to actually do depth preception but its messy.