Waymo Launches Its Self-Driving Armada
wired.com
wired.com
Back in the mid 70's there was a 'personal rapid transport' or PRT system proposed which was individual 'cars' on a rail system that were smart enough to stop at a station. The idea was that the rail line had switches where a PRT unit (4 or 8 passenger) could go into a station when called with an elevator like button. You would get in, and then it would head out on the track and not stop again until your destination station.
The idea was that it eliminated the two biggest issues with 'public transportation' which were pickup and drop off schedules that forced waits when you transferred from one line to another, and the fact that the train/bus would make all stops forcing it to be slower than a point to point bus/train.
I believe you could implement that system safely on a dedicated 'self driving car' road. And such roads would be less expensive to build than a rail infrastructure.
https://en.m.wikipedia.org/wiki/Morgantown_Personal_Rapid_Tr...
Last time we checked, there's still a plaque in the office warning about "Rotten Ralph" (his nickname). Many family stories. My sis and bro would ride unattended, monitored from the control room. They loved helping reboot the system by manually entering the hexidecimal codes.
I remain disappointed we don't have PRTs every where.
My daughter is on the autism spectrum. I think she is capable of driving herself, but at least right now, she shows no interest in it. Instead, I'm teaching her how to get around herself on a bicycle. And while I am still intent on helping her develop self-reliance, the future she will be working as an adult will likely involve self-driving cars.
I think it is very likely that the cost of self-driving cars will be higher. Middle class is already getting squeezed, and if having on-demand transportation is more affordable, it is more accessible for people. When self-driving cars have lower insurance premiums, then manually-driven cars would be left in the domain of the hobbyists and the wealthy -- driving it for fun, rather than because you need to make a living.
At the same time, I can see bicycles and electric bicycles becoming even more popular.
There was a kickstarter campaign a couple years ago documenting how roads were originally made for bikes, with the first paved roads being lobbied by bicycle groups rather than auto groups. We also have some interesting things: (1) in the early days of the auto industry, the auto industry pushed forward legislation for traffic safety, shifting the idea is that if a pedestrian is on the road, it is the pedestrian's fault, and that "roads are made for cars" (2) buying out trolly companies so people would buy more cars and (3) this idea of "roads are made for cars" has become so ingrained in our mindset that it influenced city and urban design for almost a hundred years.
I'm glad to see this is starting to swing back the other way.
Plenty of countries around the world let people start up their own bus lines to meet demand.
Speaking of PRT in the 70s, for quite some time the Morgantown PRT in West Virginia was the only functioning example of one. (I believe nowadays there are a couple others):
https://en.wikipedia.org/wiki/Morgantown_Personal_Rapid_Tran...
Throw in a standardised luggage system and lockable storage boxes at stations etc, to keep all the stuff you have around in your car, and you're getting somewhere.
Charge someone who wants to travel alone by car 10x as much as those who share and move between modes of transport
Right now you can get in your car turn off your phone and radio and go where you want in peace. I bet they build a tier systems with autonomous cars. Pay this much it will go this fast. Pay more it will go faster. There will be like 3 major companies and like the internet providers you will have to pick the lesser of all evil.
Do I think they are cool, yes. Do I like the technology, yes. Do I like where the businesses will go with this technology, no.
Moving the car seat alone makes car sharing a huge pain. And it's true now more than ever since now there's a social/safety expectation to have children in a full seat or booster seat until they're at least 10-12 years old.
Of course, if you have children, you're not using the car alone.
But yeah, 30 mins to get the car packed, and then 30 mins to clean it afterwards
https://www.amazon.com/Aramis-Love-Technology-Bruno-Latour/d...
'Yeah, they freaked me out at first too' Daniel said into the ambulance's sterile white lights, he looked back at the freshie.
'They go all the way to the hospital, right behind us, yeah?' Rick wondered if the stories were true.
'Yeah, as long as her credit card still works, it'll be right on our ass'
'What happens then? When we get to the back door?'
'Follow us right in, thank God John Muir's not a blind road and has an exit. UCSF just has a loading dock, damn things block us in all the time and we have to call SFPD to get em moved out of the way.'
'How long do they stay for?'
'In the parking lot? They'll stay until her card is cancelled or the family does something. Homelanders like y'all can't pinch two nickels, so the family'll likely never contact the card company cus they'll hound them for years about her debts.' Rick looked at her glowing green fingernails, perfectly manicured.
'After 30 days John Muir or the other hospitals just calls the car company and gets them to move. All at once. It's a laugh riot.'
As they loaded her into the bay, the shadow inched forward the exact same amount.
But then Aiden was awake.
The softness under the nylon of his car seat smelled good, like rain on Daddy's driveway. All Aiden could see was darkness and the blue and orange glow of the car. It was night now. They weren't at Daddy's anymore, but he wasn't at Mommy's house yet either.
There were noises outside, a man's voice, deep. Not Daddy's. Aiden was confused. Aiden was scared
The front door opened. The man got in, well, he more fell in. Aiden was very scared. Where is Mommy? Where is anyone?
The man smelled strange too, like Uncle Greg. That bread-like smell when Uncle Greg drank Parent's Soda.
Aiden was so scared. He just wanted to be back with his iPad dancing with Peppa! Aiden started to scream and cry! Mommy! Peppa!
'Aw fuck, therhss a kidh n thsss fuckher!' The man tried to say, but his mouth sounded like it was full of drool or sand. He smelled really bad now.
Aiden kept screaming for Mommy. Hot tears and snot.
The man turned around, his eyes were red and poofy. His hair was stringy and he was sweating a lot. He looked at Aiden but his eyes kept drifting and he kept tilting his head.
'Ssssshhhhhhh. sssalll right kiddao. Weell get ya basck hhhome. S'll call the Pawlice nn get sha basck to yer Mawmy' The man slurred a lot and got back out of the car.
The man slumped down on the side of the car and took out his phone, Aiden could see over the side of his carseat. He was so scared. Where was Mommy? Who is this man?
After a long time of crying for Mommy flashing lights came by the car. The blue and red colors were so bright. The man stood up and went over to the lights, but fell down a lot. A police lady came to the car and opened it after talking to the man.
'You ok, sugar? How are you feeling.' Aiden looked at her, this wasn't Mommy.
'It's ok now sugar, your Mommy is coming real quick, OK? Were you on the way from your Daddy's to Mommy's?' Aiden nodded and gripped the belts of his carseat tighter.
If the autonomous system can't operate in a way that passengers would be comfortable with I don't see why it should be allowed on the roads at all.
And as pessimistic as I'm being, I don't think its that bad of an idea. I just wouldn't put my money on it.
“So in autonomous mode, the car takes many precautions, so it is really safe” “Sounds good. So why do you want to let the customer drive?” “Because they will be less cautious” “I am raising your premiums”
Empty driving seems inevitable in any use-case except "my precious". Fleets, yes. Deliveries, yes. Go pick up the kids, yes.
TBH, I can see lots of problems even in existing cars for people - too many blind spots, too uncomfortable to get in and get out, more dangerous for those sitting on the driving side.
I wonder if all these problems will go away with self-driving cars, or are they necessary consequences of an optimum design to meet safety regulations.
There's really no need that every car is a >1.5 ton vehicle that can fit 5 people or more when 90% of the trips are with just one person inside.
I doubt, at typical car speeds, aerodynamics is a big factor. It's not like a jet.
Right now, people need their own car and can mostly only afford one, so they only buy cars that are at least pretty good at everything. And that's like 90% of the market, economies of scale are so strong that car companies don't really make anything but that.
Cars that drive on the freeway get a decent advantage from being aerodynamic. And right now, basically nobody wants to buy a car that can't run on the freeway efficiently.
If the self-driving car world moves car ownership from individuals to large taxi fleets, then markets open up for all sorts of specialized vehicles. Now we can make something that's super easy to enter and exit but has poor aerodynamics and high-speed handling. They might be awful on the highway, but the self-driving car taxi company will still buy some as long as they have enough non-highway ride requests and interest in the vehicle type.
That's a lot more comfortable, and already common for the rear two rows in larger cars.
I also doubt the windows will go away. Who prefers to be stuck in a windowless cubicle?
Car manufacturers today can be seen as integrators of various technologies. Most all of the driver assist tech out there today comes from companies like MobileEye and Bosch. It tends to make sense from an expense perspective to let that type of tech be developed outside of car manufacturers so the development costs can be more easily be recouped.
[0] https://www.recode.net/2016/12/21/14046496/waymo-alphabet-ho...
That's not surprising. Waymo's team goes all the way back from the 2005 DARPA challenge[1] - they've been working on it for a long time.
Their early rider program is only in Phoenix AFAIK.
> Takeaway number two: You may come in contact with these cars sooner than you think, especially if you live in one of the 25 cities where Waymo is currently testing the tech.
That said, Waymo did test the fireflies (the little bubble cars) pretty extensively in the MTV area, and I believe those had no driver at all, either safety or remote. But they also weren't classified as (basically) golf carts and not motor vehicles.
Here's Google's own statement about them having safety drivers [2015]: https://plus.google.com/+SelfDrivingCar/posts/3NuRpigxp4e
I do not believe what Waymo is launching in Arizona is legal in California.
And yeah, it's very clear that waymo is interested in moving towards full autonomy faster than the California legislature is, and so it's making advances in areas where such things are allowed.
Elsewhere you said the move to AZ was because of the difference in reporting required. While I guess technically true, I think that the real answer is broader. What waymo wants to do isn't legal at all in CA. The looser reporting required is just gravy on the ability to actually test the cars.
Some description of how your employer does business in this space can be found here: http://googletransparencyproject.org/articles/google-enliste...
The fact that a significant amount of California politicians have been bought and paid for by Google and are still resistant to adopting Google's preferred legislation on self-driving cars shows some level of how reckless Google has been here.
Alternatively, the fact that they are unwilling to adopt Google's preferred legislation disproves your premise that they are bought and paid for by Google, collapsing your entire argument.
>The fact that a significant amount of California politicians have been bought and paid for by Google and are still resistant to adopting Google's preferred legislation on self-driving cars shows some level of how reckless Google has been here.
Erm, there are about 30 conjectures in this statement. I don't know how I can have a good faith conversation with you when you say stuff like this that I can best describe as made up nonsense. Its tautological and impossible to even imagine a counterargument. Just to be clear, here's what you're assuming:
- Google's vehicles are unsafe (if they weren't what they are doing wouldn't be reckless, by definition) - Google has significant control of the CA legislature - Google is resistant to this regulation for safety reasons, and not some other reason (for example: competative advantage) - California's regulations don't harm Google's advances in areas unrelated to safety
Just as a simple counterexample: Waymo wants to beta test the infrastructure for scaling out its fleet without safety drivers. There have been 0 disengagements in the last N,000 miles (where, say, N is suitably large), and the cars appear able to drive themselves in all circumstances they encounter in Phoenix. In such a case, continuing to require safety drivers (as in CA) would be harmful to their goals, and wouldn't actually have an impact on safety.
To summarize: your entire argument that Google is being reckless is predicated on their vehicles being unsafe. If they arne't unsafe, nothing else follows. You don't trust Google. But I don't understand where "don't trust" becomes "actively assume malice at every point". That seems, strange.
Off-topic: I was surprised to see that journalists are doing Unicode art in articles now, and I don’t think that’s wise. Wouldn’t this confuse a blind reader who is listening to a transcribed version?
say '¯_(ツ)_/¯'
say '¯\_(ツ)_/¯'edit: Ah, I see your response to someone else. I'm going to try Allison.
edit: Ha! It is Allison! "Shrug". What a time to be alive!
[edit] And it "says" them exactly the same, regardless of backslash: "macron, (unintelligible)letter two underscore slash, macron".
What is says is actually this: ¯ツ_/¯
I don't know how else to run it :)
But you guessed it. I have 4 voices, Samantha/Victoria/Alex/Fred. I was using Alex. Samantha behaves as you say, all others behave as I say. (Strangely enough, now it says "Samantha (downloading)", so maybe after download it will be 'fixed'?)
For the shruggie I get "macron backslash underscore japanese in a different voice slash macron," which sucks. Too bad - I was hoping it would be recognized as a word.
I'd expect a screen reader to handle ":-)", and given that Slack has had /shrug for years, I am curious if screen readers also know how to pronounce the shrug yet.
(Is Slack screen-reader-accessible?)
There are ways to add emojis in an accessible fashion. You'd have to treat them like images and add alt text or some aria attributes to ensure the screen-reader interprets it correctly. Or, use predefined Unicode verisons.
Neither was done here and Wired should be ashamed of themselves for not even making an effort.
Markdown is not friendly to the shrug.
Emojis have been around long enough that they ought to be handling them by now anyway.
For a seeing person, there is negligible difference between "¯_(ツ)_/¯" and "¯\_(ツ)_/¯".
For either a seeing or blind person there is negligible difference between "emoji" and "meoji".
We can all make up emojis on the spot that can only heuristically be decrypted by seeing them as an image, not reading them one character at a time.
Making up a new word is still just a sequence of characters, and can be understood by any medium, whether it's written, braille, audio, etc.
There's no reasonable way to transcript emojis other than a predefined database of them, which unicode already has.
Being more adaptive than that requires offloading the responsibility of readability to the writers. If I write a blog post featuring a unique ascii emoji I made up, I can't expect a third party to handle that for blind patrons, regardless if it's a transcription service or unicode.
And transcription services should certainly recognize "¯_(ツ)_/¯" as a failed attempt at "¯\_(ツ)_/¯". Screen readers probably should too.
If they are sticking to whatever Unicode blesses they aren't serving their users well.
Also, I read provost as talking about the art in general, including a correct shrug, so I replied in that context instead of spending time on the missing \.
Nevertheless, I'm very satisfied that they're using emoticons and not emoji.
[1] https://developer.mozilla.org/en-US/docs/Web/Accessibility/A...
"Kao" is the japanese word for "face" and "moji" is "character".
That said, I don't think there's a "correct term" yet for these things. That's just what some people call them.
Tough luck, 5 million dollars to the family?
http://www.berkeleyside.com/2018/01/12/breaking-pedestrian-d...
When I write self-driving software, am I suddenly liable for all accidents made by all of those cars when my software is driving? Because the stakes there are quite a bit higher. Personally I wouldn't take the risk.
How to handle liability with a computer behind the wheel without strangling innovation isn't a solved problem as far as I know.
The liability-handling is "deep corporate pockets". That's only a risk if there's inadequate improvement.
In the case of manual driving, there is an understanding that humans might kill other humans when operating a motor vehicle and the enforcement focuses on a subset of precautions (alcohol, speed, lanes and, for professionals, continuous hours of work). For lane assistance and robot cars, there are no explicit list of guidelines yet, but there should soon be enough cases around: clearly communicating to drivers what is their responsibility; taking necessary steps (which might need professional bodies to define) when releasing something new, possibly a dedicated bureau. Cases will be more complicated because, presumably, only large corporations can handle those. The responsibility will most likely be focusing on testing practices and enforcement, not individual coders.
I suspect the closest structure will be pharmaceuticals & medical devices: it is currently acceptable to sell complex products made by large corporations that can statistically be related to thousands of people dying -- because they come with scientifically sound studies on samples, proving that they save more given a certain diagnostic. A public body defines how to prove that those help and private initiatives try to match those criteria, asking for experimental exceptions to medical practice.
Google has a habit of testing A/B on half of the world’s online population, so they might push against not trying more widely than small tests at clinical scale. There could be an interest in building a realistic simulator and testing code in there; building tracks able to replicate edge cases with mannequin -- things that actually already exist for the auto industry.
The prize is so large and the brand impact of being seen as unsafe too large for stakeholders to not find a reasonable solution.
Honestly, we have to get over this "someone must pay" mentality. If the rate of deaths falls from 1.25 million to 125,000 that will be a 90% drop in deaths. We should not be looking to punish Google or other companies for their contribution in reducing deaths because the status quo isn't good.
They don't want people thinking of Google when their cars are holding up traffic, turn out to be assholes when merging or changing lanes, or cause a fatal accident. Thus, Waymo.
In some sense, a single entity Google is an illusion so best to avoid it.
I'm also guessing it helps lure in potential leaders like John Krafcik (CEO of Waymo), because these execs can come onboard with the title of CEO rather than some VP within Google. (though, Krafcik joined Google 1 year before Waymo was founded).
TBH most of the articles I've read that mention Waymo are in relation to their lawsuit with Uber, and many of them say "Alphabet" and not "Google", and Alphabet still doesn't register as Google in my mind.
The printing press that is advertising there steals all the focus.
I'm convinced that if Waymo was a division of Google, they wouldn't be close to monetizing.
It seems Porat felt that Google would be in a better position by separating all that out so that things are more transparent to investors. If the stock price is any indication of how investors felt about that, she was right.
This doesn't mean that the Google brand couldn't still be applied to Waymo's activities, as is the case with Google Fiber, which kept the name. But it is not just Waymo that lost the Google brand in the process: Google X became just X, and Google Life Sciences became Verily.
ads by AdSense(TM)
There are many Alphabet companies: Calico, CapitalG, Chronicle, DeepMind, Jigsaw, Nest Labs, Sidewalk Labs, Verily, Waymo. Is the idea that all of these were created under the same PR-driven motivation? I think it's much more reasonable that it was a legitimate re-org of a massive company, and that organizational units want their own identity and autonomy.
Google has failed sooo hard in so many things it should be trading for $10 if it mattered.
The real effect is that, when google starts the in-car ad attack, people will act surprised.
What an odd business.
Google's search and ad network is paying (directly or indirectly) for everything. All other projects are just to keep people using their search and seeing ads on their platform.
With existing Google shareholders getting some ownership.
Alphabet would retain controlling interesting. This is pretty common but not the only way. Sometimes parent company shareholders are given shares in the new company directly. So say every share of GOOG might be given a share of Waymo Inc for example. It is not always the case that the parent company keeps control.
It can also happen the other way where a sub is public and the parent does an IPO.
The structure of Alphabet is what gives the flexibility to do this.