ML hype spawned a lot of self driving car hype, and a lot of promises which can't be delivered in any short time frame. It started with Comma.ai, but others pic

ML hype spawned a lot of self driving car hype, and a lot of promises which can't be delivered in any short time frame. It started with Comma.ai, but others picked up the full ML torch and went with it. Other companies are making steady, sure progress using the old school robotics approach of throwing sensors at the problem, Waymo, for example. Others are being reckless and trying to use vision alone (Tesla). It's a mess. However, safe self driving is coming, slowly but surely (I work at a company which produces tech for these guys). The hyped companies are in trouble, but car OEM's, partnered with companies you've never heard of, are making slow, steady progress, all the while being subject to government functional safety requirements, particularly in the US and EU. There is zero chance that a Tesla or GM car will be allowed to fully self drive, so no matter how advanced these systems are, they're sold as level 2 systems requiring driver oversight and qualify as cruise control in regulations. Today, we have full autonomy in some truck routes (only as proofs of concept), in ship yards, parking shuttles, mining equipment, quarry trucks, etc, places where the problem domain is more constrained. Generalized self driving is a ways off, but by the time you can make use of it, it will be safe, it just won't come from Tesla or Cruise or Uber.