/biz/ - Business & Finance

>>58632655
>He says this is because narrow artificial intelligence is not capable of handling an edge case it has never seen before

it's a short sighted argument because he's not taking into account that this is an active area of research so they will be training the cars even better in the future. plus now they have a lot more training data in the last year than they ever had before as the amount of miles their cars are self driving has skyrocketed. 10 years ago we had pretty much no self driving cars on the road and they were only self driving at research facilities. now we have millions of cars with the capacity to self drive and they have recently crossed the threshhold of driving over a billion miles. what happens when these cars have trillions of more miles driven, or hundreds of trillion. and who's to say that future self driving cars won't be able to see things better and more accurately and understand the world better. who's to say that in 10 years we wont have self driving cars that are orders of magnitude smarter than they are now.

also a lot of times when people say "Never" in the technology field they are proven wrong eventually. these cars are not an impossible technology, they already exist and are already quite good at a lot of things. they will for sure get much better.

>>58304159
>Everyone can see FSD in action on youtube, it doesnt work, let alone as a fuckign robotaxi
There is an inflection point for miles driven with FSD so apparently somebody likes it