> In any case, the Mercedes system is what's known as a conditionally autonomous system.
Yes, but it's significant progress if it reaches consumer hands. This sounds like the typical Waymo marketing of "you can't get to Level 4 through Level 3". I bought into this at one point, but Waymo and Cruise have demonstrated that going straight to Level 4 with $100k+ (probably much more than that) of sensors and corresponding maintenance doesn't scale.
If other companies thought they’d hitch on to Waymo after they started giving rides in 2017, then that’s on them. Because Waymo has been pretty clear that their scaling is going to be methodical and not instantaneous.
Waymo won't keep a monopoly on this forever. There are still high barriers to entry, but they will fall sooner or later. If there is money to be made, competition will come.
I'm not sure about barriers to entry lowering soon. First off, this is an incredibly hard problem no one else seems to have cracked (and even Waymo has a long way to go). Second, they could help influence regulation that requires extremely high safety records, mandatory sensors, etc. which could make it harder for others to compete. It's all very early though.
No OP but there were multiple headlines of Waymo incidents just recently
- https://techcrunch.com/2024/02/13/waymo-recall-crash-softwar...
- https://www.theverge.com/2024/2/7/24065063/waymo-driverless-...
In any case, the Mercedes system is what's known as a conditionally autonomous system. It can drive itself under specific conditions, but a human is still involved in operating the vehicle as soon as it exits that design domain. Waymo (and Cruise) are fully autonomous systems. Humans are not involved in the control loop and the vehicles will stop if they leave their operational domain. Tesla FSD is not an autonomous system at all and requires a human to always be in control.