It just makes sense for CEO’s to make these statements publicly so consumers understand. There are Tesla drivers out there who think autopilot means the car can drive autonomously.
Frankly, it was stupid for Tesla to call that feature “autopilot”.
Either way it will take a long time to replace the national fleet with the new technologies but we should see accident rates begin to decline steadily.
I continue to stand by my 2017 prediction that we won’t have self-driving cars before 2024, and I continue to believe it won’t be until at least the 2030s.
Are autonomous cars just five days away? I doubt it but they have something they want to draw attention to.
I still maintain that one first use of this technology will be in amusement parks and airports. With a self-driving shuttle, remote parking lots offer the little bonus, a bit of novelty, of a robot ride to the main site.
I was driving in a snowy hell in South Bend for the first three months of 2019: ice, melting ice, slush, blizzard, almost too snowy to drive at all (got stuck in a snow bank–plow guy just happened to be working nearby–thank you!), rain on ice, slippery powder–you name it. All with cars driving fast and tailgating me all the time (what is up with South Bend drivers? I was never not tailgated!).
There is really a high amount of skill and experience required to drive in all of these conditions. Plus, often you can’t see the lines on the road, and sometimes you can barely see anything at all.
I thought about self-driving cars at the time, and my conclusion was, “Wow, I bet the snow driving module for these things is going to be some pretty heavy-duty AI.” As in, “Actually, it’s impossible.”
It probably can be done. With AI close to Strong AI level, a painstakingly developed library to cover every type of driving condition, and tons of redundant markers buried in the roads, etc., to help guide the bastards.
Close enough isn’t good enough. Experienced drivers such as myself do get caught in snow banks and worse because of tiny errors of judgment, or circumstances are not what they seem (black ice, etc.). Getting caught in a snow bank is one thing, but getting someone killed isn’t forgivable. Really, really good AI will get people killed in such conditions until it’s breathtakingly, amazingly good. That’s not going to happen any time soon.
or they could do the easier part first and not include snow.
where I’m at, we see 2 or 3 inches of snow for 1 or 2 days about every 3rd or 4th year.
I dunno how many places or what % of the market is similar, but there must be a substantial portion of the market that will never need to drive in snow, and likely a decent part of those that do would still want a self driving car for the summer months
Yep, but that wouldn’t be the vaunted “Level 5.”
The thing is also: it’s not just snow. It’s that and a handful of big problems like it and a 1,000 other small ones that people have thought of, and a 1,000 others that people won’t think of until something bad happens to make them think of it.
This happened with regular ol’ cars as well. Pretty much anything designed before the year 2000 is a death trap, and the things were on the roads for decades before people even thought of, you know, lap belts (or they at least didn’t put them in).
Similarly, people in the year 2100 are going to chortle and say, “Can you believe they put self-driving cars on the road without even having X?!”
GM’s Cruise unveiled its first self driving car: GM’s Cruise unveils its first driverless vehicle
Optimistic take is “Self-driving cars are almost here!” Jalopnik is reporting that this is not a concept and that it is ready for delivery. Pessimistic take is “[CEO Dan Amman] did not say when the vehicle would go into production nor how many the company planned to build. It has not been approved to drive on roads, and it will require extensive testing before this is granted.”
Thanks commasense. My first link was supposed to go to that BBC article. Carelessness caused me to link to something completely irrelevant. I’ll ask the mods to break my link.
That Jalopnik article features a picture of the Cruise running a red light. 
"It’s looking like we might be feature complete in a few months…
… feature complete just means [the car] has some chance of going from your home to work without intervention. That doesn’t mean the features are working well.”
“Musk further specified that the FSD software will provide an ‘above zero’ chance that your car can successfully guide itself from one place to another.”
Somehow, that’s not filling me with confidence.
A VW executive working on AV is saying L5 may never happen, but they do think they can make L4 workable.
Yeah but Troutman, that’s the best we can say about you.
Looks like it’s making a right turn on a green to me.