Not proven does not necessarily mean "wrong track."
Vision alone has not proven itself to attain L3 so far, so it takes Tesla to prove that it will.
The thing is, how long would you decide the race of sensor fusion vs vision-only will finish? Since no one has finished commercializing L5 yet, so it's hard to decide a winner. Yes, speculators can guess who will win, but the result won't reveal itself until the finish line.
The rationale for vision alone is simple: people can walk with legs, birds can fly by flapping their wings, fish can swim with flippers so just imitate nature and you are on the right track!
The problem is: Cars don't walk with legs, jet airplanes don't flap with wings, submarines don't flipper...
We found better than legs for the car to move. Better than flapping wings for a jet to fly, and better than flippers for a submarine to swim.
That's the same with Autonomous Driving: LIDAR is better than vision to detect obstacles, especially in the dark:
A Tesla Cybertruck in Full Self Driving mode crashes head on into a light pole. Is this a joke? No, sadly, it is not.
cleantechnica.com