Anyone ridden in a driverless taxi?

I did for the first time yesterday in a Waymo navigating the streets of San Francisco. Blew my mind!

I thought it would be slow and timid but it navigated very busy streets confidently and safely.

Lane changes, other cars coming at us head on (in our lane), going around cyclists and e-scooters who were dodging parked cars, pedestrians stepping into the street without looking......it handled all of it with ease. There was one situation where our car needed to turn right but it paused for a few seconds. I thought it was a glitch at first but a cyclist flew by us on our right-hand side from behind, then the car made the turn. It saw everything way in advance.

I was very impressed with the technology.

Oh, and I should probably add this considering some of the comments I expect below:

 

You sir, have more of a death wish then I do & I do LOTS of stupid shyte, Just ask my wife!!! I used Uber for the 1st time when we went to Det. for Bruins vs Wings game. No way was I gonna rent a car in Det, pay $40+ a night for valet hotel parking. It worked out great, still not interested in driverless car. Like the 3 stooges episode they had the headless driver!

 

I was definitely nervous hopping in but most of those nerves settled as soon as I saw how it handled various situations.

The thing was a better driver than my wife!

 

I'm sure that AI cars are fine..... in 99.99% of the situations they encounter.

I'm in the software industry. What bothers me about these things isn't that they can't handle normal situations, it is what we call the 'edge cases' - situations that the programmers didn't think about when they developed and tested the software.

If the software does something wrong (and lets be real, in a high traffic environment, doing the wrong thing can be catastrophic) in one of these edge cases, it will do it wrong EVERY SINGLE TIME until the code gets fixed.

This is a very geeky aviation video, but it goes precisely to the point I am making with a Boeing system on the infamous 737Max called the MCAS. In this case, that edge case cost the lives of a lot of people:

https://www.youtube.com/watch?v=DcB31RgeL50

 

I agree, that's definitely a concern.

There will always be accidents involving these things no matter how good the software is. Just like there will always be accidents with human drivers.

I guess it comes down to how good is the robot car vs the human driver?

 

I'm a tester and can't tell you how many times I found a discrepancy that is almost always described as an "edge case". You also have the issue of ethical programming - if the vehicle is going to crash and most likely kill the occupants, but the only way to save them is to swerve into a bicyclist or a group of school children, it's up to the programmer to decide what will happen with little/no existing ethical standards and essentially no regulations (at the moment).

Honestly, even though it's 3D vs 2D, I think programming flying vehicles is probably easier than on-road vehicles. I believe e-VTOLs may be a better future than electric cars

 

I never call them Waymo (dumb name, by the way). I've always called them Johnny Cab's.


https://www.youtube.com/watch?v=eWgrvNHjKkY&t=2s

 

God bless the folks who choose to write this code. They are serious and diligent professionals who take what they do with the gravity it deserves. But they are human, and mistakes will always happen.

I don't think I could live with the consequences of a mistake that cost innocent lives.

 

Waymo's "machine learning", "neural network", etc. is done using Los Angeles drivers as the training model

 

Jake, You have quite the sense for adventure! The technology part of it is 'numbing'. I'm thinking I would give it a little more time before trying a ride in a driverless cab, like till the software glitches were reduced to almost nothing. Someday when I feel like gambling, maybe I'll try it then! Thanx for the report though.

 

I tried when I was in Scottsdale a couple of months ago for work. But the wait time was an hour for the next available. I didn’t have the patience and just took a regular Uber.

 

What caused the crashes W/MCAS was LACK of third world training. I was called out by some 1/2 @$$ pilot on the MCAS when it 1st came out on a RW forum, turned out I was correct. No crashes by FAA, EASA operators because they were trained to turn off the MCAS when it thought the A/C was in a stall configuration. The FAA allowed Boeing to sell those to SW Airlines knowing there was a software issue because Boeing was facing a huge $$$$ penalty from SW for late deliveries. Boeing has really gotten away from their quality too. The B787 in production has NO DOT Drug & Alcohol Testing for the employees per 14 CFR Yet the FAA has NOW mandated that foreign repair stations will need to comply W/random Drug & alcohol testing finally!!https://www.faa.gov/newsroom/new-fa...ohol-testing-foreign-repair-station-employees

 

I'm in the software industry as well and, while I take your point about edge cases, do you think human beings are fine 99.99% of the time? I think not. I think we would be generous in saying drivers I typically encounter on the road are any better than 80%.

 

Regardless of rather they are 'right' or not, human drivers are able to adapt to strange situations on the fly and are incredibly safe overall. Driverless taxis don't exist because humans are inherently dangerous, they exist for economic reasons.

 

Last ride in a Waymo over the holidays, we were stuck behind someone who had double-parked in San Francisco and Waymo wouldn't go around it. We just sat there. My wife finally reached over and tried to honk Waymo's horn and all hell broke loose. All these alarms go off and the customer service calls in (after another several minutes) and tells us not to touch any of the car's controls. We said, well then get this thing moving! After a long delay, it finally pulled out and around. Have had other issues with Waymo, too, like it won't drop off at my family's place. Wants to leave you with a good walk ahead, despite it being an accessible street that other ride sharing companies, along with residents, have no problem with. Not impressed.

 

I think humans may be able to adapt, but humans are thinking less and less. And driverless vehicles are a thing because technology is growing and that’s another way to let us be more lazy. And from an economic standpoint, I’d expect the insurance industry to be against this, as there will be less accidents with more computer drivers.

 

Large corporations carry very robust insurance anyway. And the profit margin would likely more than cover the difference, if there was one. My issue with this, and lots of other technology, is if it’s just driving up the profit margin while putting good people out of work, I’m not a fan.

 

I agree automation is continuing to take jobs, but computers continue to do things more consistently than humans.

 

And more automation is coming, whether anyone likes it or not. People are going to have to learn to adapt and survive if their livelihood is taken by automation. Easier said than done, especially for older folks who've been doing one job their entire life

 

As a 70 year old I’ll chip in, it depends on the design, a good design makes life easier. A bad design makes life harder. Think crawl control as a good design and tax software like going to a dentist, lol.

 

As someone who filed taxes via the paper form, tax software is way easier.

 

I'll agree with that. I'm not a fan of cramming tech/automation into something just because it can be done. I prefer the KISS (keep it simple, stupid) method. I have, however, tried to automate things at work when it makes sense. Such as completing a repetitive computer task that just sucks up someone's time.

I don't understand the push for driverless cars given the level of complication - there are just so many variables that make the programming... problematic. Like construction, unpainted or poorly painted road lines (mostly due to construction), vandalized signs, animals, weather, and the biggest variable of all - other human drivers. I could be wrong, but I think the only way for it to really work well is for all vehicles to be connected (think internet or satellite) but to have every vehicle on the road connected would take another 50+ years, simply to wait for all the unconnected vehicles to die off. This also assumes cross-company connections becoming normal in the next 20 years. I just don't see it happening. Maybe AI changes things and they become better without needing to know where all other vehicles are and what those vehicles are going to do.

I still think moving into the air is a better option. Autopilot automation has been tested and advancing for a very long time, the shortest distance between two points remains a straight line (for now), and you can move faster through the air. Maybe I'm just stuck on The Jetsons, although I doubt vehicles will ever be able to fold up into a briefcase

 

This made me think of this post.

https://youtu.be/j4JNn33BDG0?si=M1B1YXbucvyTgpFN

 

This post gives me hope for the future!

 

Haha, I've never seen that before. When I first read about UBI, it was in a book called Physics of the Future. Michio Kaku believes that with the rise of automation, UBI will be a necessary evil because there just aren't going to be enough jobs left for people to do.

I'm not sure how I personally feel about that but I guess I can see the logic. It was interesting to see a presidential contender bring it up a few election cycles ago (Andrew Yang). Not sure there was a lot of support for it then, or even now. I don't think we're anywhere near an automation tipping point yet.

 

The thing about aviation automation is that it requires highly trained pilots managing and monitoring those automated systems. Which is precisly the scenario driverless cars are trying to replace.

and most of what you hear about 'ai' is a bunch of hype that is nowhere close to delivering on the promises you hear about in the media.

 

.
...^^^... One of my foreign friends (she is like a surrogate daughter) that I helped through school from Indonesia does a lot of that coding & whatever for Cruise in SF.

Super brain grad from GA Tech at 20 and Columbia Masters @ 22 , MIT PhD at 24 1/2, then to Tesla making $330K + stock then to Cruise making a fortune.

Still a kid at 29 1/2, she makes a ton of money BUT she is kind of on call 24/7/365 and burnout is often and gets lots of calls in the middle of the night.

She has had to take several 3 month breaks to keep from going nuts ... and she goes back to Indonesia to visit her family for R&R. I could not do her job.

She recently bought a small place in SF for $1.4M and it's not very big.

PS ... She tells me the stories about the kids coming up with a new app, program or inventions and selling them to angel investors and becoming rich overnight. Many of those investments fail, but the whiz kids leave with millions in their pockets.

 

He had a good podcast a few years ago and touched on a variety of interesting topics.

 

LOL first thing that showed up on youtube when I search it

https://www.youtube.com/watch?v=n0YBAzV2bzY

 

Michio Kaku or Amdrew Yang?