To clarify, that's not to diminish what these companies (and other similar ones) are doing -- this is a fundamentally unscoped problem that is being solved. Above all of them, what Waymo has accomplished is truly shocking, and a testament to the power of being about to throw hundreds of engineers for close to 20 years at a complex set of problems. Lots has been written about this elsewhere, but highway driving is fundamentally "easy" compared to the problem of "where do I drop someone off at a shopping center with a busy parking lot, and how do I get there". There are uncountable edge cases in the best of conditions, even before you consider the nuisance of having to interact with humans.
What Tesla has built is also amazing, but not even remotely at the same scale. It's incredibly useful in a variety of conditions (especially highways) and I depend on it almost daily, but it is absolutely terrifying in less controlled conditions and I can't imagine using it in any sort of unsupervised way (again, except tightly controlled conditions) in the next decade. Even the "FSD Beta" is not solving the same problem.