I can see where you're coming from.
That said, my wife is *in* autonomous tech and EV. She sits on the board of and advises a few startups and the M-City initiative at the University of Michigan (pretty cool, if you want to click-through: https://mcity.umich.edu/). I am her test audience before she goes and speaks at conferences -- this includes South By a few years ago, shout out to Austin -- or publishes papers or whatever, so I have that spousal "you have to know about this shit because your wife is tits deep in it" thing going on.
In any case, the consensus in the industry is that there were two paths that companies took a few years ago.
Tesla's path was, more or less, a bet on cameras. This was cheap and quick to deploy, and allowed them to jump out to a head start. Consider that you can get a 1080p backup cam for like $20 on Amazon; imagine buying at scale. Anyway, consistent with Elon's patterns, he marketed the hell out of this, including the "vision" you mentioned above. However, it appears to be bearing out that a system founded upon cameras is easy to quickly build up to Level 2 or 2+, but then you hit a wall. This is why Tesla is no longer seriously advancing and is flailing to put together a new hardware stack in hopes that it can move the ball forward. It is unlikely that Tesla will ever reach Level 4 or 5 with its current approach and the consensus is that Tesla is too committed/down the path of to its current approach to pivot and recreate the software side for a more fulsome system.
The majority of the industry uses some combination of cameras, radar, lidar, etc. At first, many companies were well behind Tesla because of some combination of cost (prohibitive to deploy enough sensitive) and the software side (integrating the various sensor systems. Now, the cost of the hardware is plummeting (example: https://www.bloomberg.com/news/articles/2022-10-25/lidar-sensor-price-plunge-leads-to-orders-from-gm-volkswagen). Also, the physical size of a lot of these systems is shrinking and they're becoming easier to integrate. I wish I had my wife's slide deck showing the change, with the pic of the lidar from a few years back that looks like an old school police siren on top of the car to the little sensors that are coming online now. Anyway, between decreasing hardware costs and increasing hardware availability for the various sensor systems, and realizing the gains from all the investment that has gone into integrating multiple systems for higher-level automation, other players are now leap-frogging Tesla. This is why you see that MB just got their Level 3 approved and GM's Level 3 successor to Super Cruise is coming soon, why Ford's Blue Cruise is considered a leader now, etc.
Trying to think of a good analogy, but maybe it's like a sprinter racing a distance runner over a half-marathon. Tesla jumped out to a big lead but, with its camera-based approach, is not equipped for the long run. Meanwhile, the majority of the industry has been plugging along, saw Tesla jump out to a big lead the first 100 or 200 meters, but is now passing it.