Cake
  • Log In
  • Sign Up
    • I find myself not really caring that much about the self-driving car race. Not because the technology isn't cool, it is, but because I'll likely never be able to enjoy it here in Malaysia. The three companies in the race don't have a big presence here. Uber lost to local competitor Grab and left the country, Google doesn't even sell it's Pixel phones here, and Tesla doesn't exist in Malaysia.

      Having said that though, to answer your question, I'd probably go for Uber to win too. They already have the ride sharing infrastructure on a mostly global scale. Even if Tesla or Google were to win the race to get road-ready self-driving cars, they'd then need to win the ride-sharing market too. Uber is already half way there and just needs to figure out the self-driving bit.

    • Not sure how far into the future you're thinking of. As the NYT reporter who covers Silicon Valley said, Facebook won't be a thing in ten years.

      So my guess would be operator(s) that either don't exist right now, or who we haven't heard of.

      Running a taxi business would be a whole new venture for Tesla. Worth their time? Or better for them to sell/lease their vehicles, thus staying in their lane, as it were. :-)

    • I guess my thoughts are on the law of unintended consequences and who the less obvious losers are going to be as a result of this technology shift. For example, what about individuals with disabilities who need human assistance to get into a taxi? Will we require another Gang of 19 protest to make assistive technologies a requirement, or will transport options shrink or become unaffordable for the disability community?

      Also, what happens if you’re driving your non-self driving car and get hit by a self driving car? Are your insurance rates going to be raised more than if another human driver had hit you?

      It appears that we’ve migrated to a gig economy, more so amongst Millenials. Will there be boycotts when Uber drivers lose their jobs?

      There will obviously be a transition period where you have both driver and driverless taxis: it will take time to build up a driverless fleet. Will Uber drivers exploit that vulnerability, organize and negotiate a mandatory number of human drivers to remain—or at least negotiate for a better exit package for not going on strike during the transition? The maths of game theory escape me, but there is an enormous advantage for the Uber drivers to organize now so that they are in a better negotiating position when the robot overlords assume their rightful place.

    • Uber is reckless and doesn't have the engineering needed to pull it off. Tesla is hamstrung by the tech stack they chose (visual only, no lidar). My money would be on Google. Not only because of the tech, but because of the approach: solving only the core problem, and letting the automotive industry partners take care of the boring parts (design and manufacture of the actual vehicles).

    • jpop, you always have great technical insight so I pay close attention when you say something like no lidar. On the other hand, Musk makes a pretty good case that it's in the wrong spectrum and the real problem that has to be solved in any case is visual recognition, no? And he argues he has radar, which can see past obstructions, and ultrasonic sensors.

      He said he realizes if he's wrong he'll look like a fool but he's quite sure he's not. I thought this was a pretty good dive into why he may not be wrong:

    • I think there are multiple roads to success here. In other words, each company can use the strength of it's history and resources and go from there. I think all of these companies have more than enough money and resources to be successful. Sure one might take an early commanding lead but it won't last. Of course I could be wrong.

    • Yesterday a neighborhood friend was frolicking on our electric scooter out in the street while I took pics.

      Along came a Waymo, like they do a dozen times a day, and I noticed that my wife and Addy's parents started yelling for her to get off the street and onto the sidewalk as any parent would when a car was coming.

      What was different this time is they started shouting that it was a self-driving car and so it may not see her. It made the alarm level even higher than usual. Interesting consumer perception.

      I am used to Waymos and had no worries, but as Addy eventually headed for the sidewalk I noticed the Waymo car never slowed down like most drivers would have with a kid on a scooter in the area. I felt myself getting a little bit mad about it.

      Last night I noticed a bicyclist with no lights and a Waymo was about to pass it from behind. I watched very closely to see if the car would slow down or move to the left to give more space to the bike, but no. Most drivers would have as a precaution. I didn't feel too good about that either.

    • You should contact the company and ask why? What is their policy and reasoning. I'm sure it's available and they want to make people accepting so they would listen to your quiry with interest.

    • Good idea. The funny thing is California has this on the Department of Motor Vehicles website:

      Passing a bicyclist that is in the travel lane at a safe distance may 
      require changing into another lane, passing safely and quickly, and 
      returning to your original lane leaving room between your vehicle and 
      the bicyclist. When you cannot change lanes to pass a bicyclist, allow 
      at least 3 feet between your vehicle and the bicyclist. If you are 
      unable to pass within 3 feet of space, pass at a safe speed to not 
      endanger the bicyclist.

      I was paying very close attention and I'd be very surprised if my judgement was wrong. I thought the Waymo was closer than 3 feet to the cyclist and didn't slow down.

    • Sure, it may turn out that Tesla is right, their approach is not entirely without merit. But cheap, solid state lidar is just around the corner, and pure machine learning approach that Tesla is pursuing is risky. I would be much more at ease with Google's rule-based, street-scanned supervised machine learning. It may take longer to deploy widely, but rush is not what we need in deploying of this kind of tech. Tesla might get there first, but would you personally risk using their pure machine learned autopilot, knowing that at any time it might encounter a novel situation that might make it confuse a tractor trailer for clear skies?

    • I am used to Waymos and had no worries, but as Addy eventually headed for the sidewalk I noticed the Waymo car never slowed down like most drivers would have with a kid on a scooter in the area. I felt myself getting a little bit mad about it.

      This is a really interesting point.

      When I'm driving in a residential area and there are kids in the street, I slow down not just to be safer, but also to communicate to everyone around, "Hey, I see the kids in the street and I'm being careful. Don't worry!" I also give cyclists lots of extra room for the same reason — to help them feel safe.

      Those Waymo cars probably do see pedestrians and cyclists and are probably perfectly capable of not running into them, but since they're not communicating that fact like a human driver would, it makes people worry. Waymo should fix that!

    • I was paying very close attention and I'd be very surprised if my judgement was wrong. I thought the Waymo was closer than 3 feet to the cyclist and didn't slow down.

      Truth is not only the car's reactions need to be accounted for, but also the pedestrian or bicyclist may unintentionally swerve and then the Waymo has no error margin. I think roads aren't safe shared with such vehicles and they should test them in controlled environment, not risking real human's safety, for their faster advancement of profit.

    • We'll have to decide what reasonable risk is. How many deaths are acceptable per so many miles of driving. If it's twice as safe as humans would be accept it? 10X safer? At what point should we allow it on the roads? If one company's approach is 100X safer than a human and anothers is 120X safer, should we accept 100X? Questions that we'll have to decide on.

    • I listened to a huge debate about this. Some people, like Elon, think the number we need is to show that self-driving is twice as safe and then people will accept it. The other point of view I know some people have is no, it has to be like airlines. There will be rage if there are any accidents where self-driving is at fault.

      I was in Elon's camp until @jpop mentioned the Tesla that didn't see the tractor trailer in clear skies and then I was like, hmmm, maybe it's twice as safe as other drivers, but it's not as safe as me. What is the number? 80% of American male drivers believe they are above average?

    • I certainly would trust the car over my own driving and in 32 years of driving I've never hit another person. I think I trust technology and science much more than the many mistakes I know I make while driving. I think the Dunning-Kruger effect applies to drivers. We get tired, we are distracted, we are slow to react, we don't know how to do slides and so on. Humans simply suck at a lot of stuff and computers can do it better for the most part. I've watched a video of a robot driving a street bike around a race track and humans can certainly outperform the robot rider but even that won't last. It'll be something to see a robot back a street bike into a corner and pass a pro. I want autonomous driving yesterday.

    • Related fun fact to this conversation:

      Today in 1912, construction of a stadium in Brooklyn was announced for the Trolley Dodgers.

      Apparently, it was a regular life and death struggle in early 20th Century Brooklyn to avoid getting hit by a trolley.

      The DC powered motors (unlike the AC models)  jerked and bucked when started.  Trolleys could jump 20 to 50 feet, posing threats to  pedestrians and horses alike.

    • i'm not sure who will "win" this, but my personal preference is that they ALL lose. talk about an answer for a question that didn't ever need to be asked!

      the self-driving car is proof of the failure of efficient transportation. the last thing the environment needs is more complicated ways to have personal transport. self-driving cars will NOT help the environment; if anything they will harm it, as the infrastructure needed to support this will be more energy-intensive. i could possibly see self-driving vehicles for mass-transit, but no one needs self-driving vehicles for personal transport. except that maybe the future cab companies want to cut back on labor costs, to serve those who want personal transport occasionally? it's a disaster for the environment, tho...

    • Hi Doug, and welcome to Cake. 🙂 What if it turns out that self-driving cars become significantly safer than human-driven cars? Might it be one of those things where it's rude to have a self-driving car in the beginning but later if they are safer it becomes rude not to have one?

    • it won't happen. and even to get as safe as human-operated cars, it will be prohibitively environmentally costly. the planet simply cannot afford to continue squandering its resources in its present manner. to have a complete infrastructure overhaul to support driverless personal vehicles would be a massive undertaking, further destroying our ecosystem. how safe is *that*? better to invest in mass transit systems. that would increase safety *and* reduce the environmental impact of human beings.

      my estimate is that the earth cannot support more than 1-2 billion people at its current use of resources. mother nature is a harsh mistress, and she *will* self-correct... *no* species is immune form the j-curve...