4 ms·
Just FYI "basically requires zero intervention" is nowhere close to what's needed for self-driving. It needs to be probably 1-2 orders of magnitude better than
by sorcerer-mar 1y ago
Just FYI "basically requires zero intervention" is nowhere close to what's needed for self-driving.
It needs to be probably 1-2 orders of magnitude better than what the vast majority of people would actually experience as "zero interventions ever".
One intervention every 80 miles ("basically zero") is 178 interventions per year for the average American driver. Let's say conservatively that 10% of those avoids an accident, you're looking at 17 accidents caused by each driver per year, never mind the accidents they're just victim to due to misfortunate of being near someone else's 17 annual accidents.
- homefree 1y agoIt’s still supervised, the Austin robotaxi is running unsupervised with newer software in a more controlled rollout. For supervised minimal intervention is fine and 99% of the time I’m just sitting there watching. It’s still extremely valuable and useful and something nobody else is close to. It makes driving a lot simpler and safer. I already think it’s better than most people and would rather have a driver use it than not if I’m a passenger. Unsupervised will be amazing if they can pull it off generally and will have major market effects, but obviously it’s harder for the reasons you describe.
- sorcerer-mar 1y agoI'm sorry for the direct snipe, but: what the fuck are you talking about?