4 ms·

Why wait till it's ready?

uniqueid 6y ago
More seriously, the nice thing about this "prediction" is that it's so close. So I can go on record and say: Teslas will not be fully self-driving "by the end of the year" nor even by the end of next year, nor the year after that. I doubt we'll see one in 2030, for that matter. If a car isn't smarter than a human, it will kill people in ridiculous circumstances. Reaction time can be an order of magnitude better than human, but you can't release a car that occasionally drives over a baby for no reason. If Musk actually claimed otherwise in earnest, he's either having another episode, or maneuvering to fix an issue with PR or stock price.
- grecy 6y ago
  > you can't release a car that occasionally drives over a baby for no reason. I disagree entirely. We have planes that occasionally kill people, we have x-ray machines, microwaves and gas BBQ's too. And of course, 36,000 people are killed on the roads in the USA each year [1]. Any self driving car doesn't have to 100% perfect, it only needs to be better than human drivers, and it will be saving lives. Yes, the "bad press" will be painful, but in the long run less people will die as a result, so it's a net win. [1] https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in_U.S._by_year#Motor_vehicle_deaths_in_U.S._by_year https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in...
  - jocker12 6y ago
    I am not sure you are looking at the correct numbers. according to NHTSA – https://www-fars.nhtsa.dot.gov/Main/index.aspx https://www-fars.nhtsa.dot.gov/Main/index.aspx there are 1.18 fatalities per 100 millions miles driven. That means, if an individual drives 15.000 miles per year, that individual will face the possibility of dying in a fatal crash as a driver, passenger or pedestrain, once in 6666 years, so the cars and road system are extremely safe as they are today. Most of the self driving cars developers recognize this like Chris Urmson in his Recode Decode interview – “Well, it’s not even that they grab for it, it’s that they experience it for a while and it works, right? And maybe it works perfectly every day for a month. The next day it may not work, but their experience now is, “Oh this works,” and so they’re not prepared to take over and so their ability to kind of save it and monitor it decays with time. So you know in America, somebody dies in a car accident about 1.15 times per 100 million miles. That’s like 10,000 years of an average person’s driving. So, let’s say the technology is pretty good but not that good. You know, someone dies once every 50 million miles. We’re going to have twice as many accidents and fatalities on the roads on average, but for any one individual they could go a lifetime, many lifetimes before they ever see that.” – https://www.recode.net/2017/9/8/16278566/transcript-self-driving-car-engineer-chris-urmson-recode-decode https://www.recode.net/2017/9/8/16278566/transcript-self-dri... or Ford Motor Co. executive vice president Raj Nair – “Ford Motor Co. executive vice president Raj Nair says you get to 90 percent automation pretty quickly once you understand the technology you need. “It takes a lot, lot longer to get to 96 or 97,” he says. “You have a curve, and those last few percentage points are really difficult.” Almost every time auto executives talk about the promise of self-driving cars, they cite the National Highway Traffic Safety Administration statistic that shows human error is the “critical reason” for all but 6 percent of car crashes. But that’s kind of misleading, says Nair. “If you look at it in terms of fatal accidents and miles driven, humans are actually very reliable machines. We need to create an even more reliable machine.” – https://www.consumerreports.org/autonomous-driving/self-driving-cars-driving-into-the-future/ https://www.consumerreports.org/autonomous-driving/self-driv... or prof. Raj Rajkumar head of Carnegie Mellon University’s leading self-driving laboratory. – “if you do the mileage statistics, one fatality happens every 80 million miles. That is unfortunately of course, but that is a tremendously high bar for automatically vehicle to meet.” min.19.30 of this podcast interview – http://www.slate.com/articles/podcasts/if_then/2018/05/self_driving_cars_are_not_yet_as_safe_as_human_drivers_says_carnegie_mellon.html http://www.slate.com/articles/podcasts/if_then/2018/05/self_... or Elon Musk himself - "On average, a person is killed in a traffic accident in the United States once every 100 million miles. Elon Musk says Tesla’s Autopilot is half as likely to be involved in a collision as a human driver. That would suggest that somewhere around the 200 million mile mark someone will die as a result of an automobile driven by a machine." - https://www.teslarati.com/autonomous-cars-make-us-worse-drivers/ https://www.teslarati.com/autonomous-cars-make-us-worse-driv... What you are using is a fallacy, emotional statement done by self driving cars developers and enthusiasts in order to make people think by adopting this technology they will be part of a bigger better future, by doing essentially nothing."
    - grecy 6y ago
      What I'm saying is that self-driving cars don't have to be perfect, they just have to be better than human drivers. Once there are less deaths with self-driving cars that human-driven cars (however you want to measure that) then it makes sense to use it.
      - uniqueid 6y ago
        > they just have to be better than human drivers. That's where the difference in our outlook is: The way I see it, people can stomach the existence of ridiculous accidents (ridiculous in the sense that the details are "absurd", not just "very grim") when they can blame a human. I don't think many people will tolerate even occasional "self-driving" accidents if the circumstance show a human would have recognized the danger. For a self-driving car to be fit-for-purpose, it can't run over kittens at low speeds on an empty road - even if they're very small kittens that don't exceed some motion-sensing threshold. That requires common-sense reasoning.
        grecy 6y ago
        I think you're absolutely right there will be push-back if the self driving cars crash is stupid ways, but again, it's all about the numbers. Planes still fall out of the sky for stupid reasons (maintenance errors, mostly) yet the numbers say they're very, very VERY safe, so people use them. Once the facts show self driving cars are safer than human driving cars (even if they sometimes crash in stupid ways), there will be no holding back.

More seriously, the nice thing about this "prediction" is that it's so close. So I can go on record and say: Teslas will not be fully self-driving "by the end of the year" nor even by the end of next year, nor the year after that. I doubt we'll see one in 2030, for that matter. If a car isn't smarter than a human, it will kill people in ridiculous circumstances. Reaction time can be an order of magnitude better than human, but you can't release a car that occasionally drives over a baby for no reason. If Musk actually claimed otherwise in earnest, he's either having another episode, or maneuvering to fix an issue with PR or stock price.

grecy 6y ago
> you can't release a car that occasionally drives over a baby for no reason. I disagree entirely. We have planes that occasionally kill people, we have x-ray machines, microwaves and gas BBQ's too. And of course, 36,000 people are killed on the roads in the USA each year [1]. Any self driving car doesn't have to 100% perfect, it only needs to be better than human drivers, and it will be saving lives. Yes, the "bad press" will be painful, but in the long run less people will die as a result, so it's a net win. [1] https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in_U.S._by_year#Motor_vehicle_deaths_in_U.S._by_year https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in...
- jocker12 6y ago
  I am not sure you are looking at the correct numbers. according to NHTSA – https://www-fars.nhtsa.dot.gov/Main/index.aspx https://www-fars.nhtsa.dot.gov/Main/index.aspx there are 1.18 fatalities per 100 millions miles driven. That means, if an individual drives 15.000 miles per year, that individual will face the possibility of dying in a fatal crash as a driver, passenger or pedestrain, once in 6666 years, so the cars and road system are extremely safe as they are today. Most of the self driving cars developers recognize this like Chris Urmson in his Recode Decode interview – “Well, it’s not even that they grab for it, it’s that they experience it for a while and it works, right? And maybe it works perfectly every day for a month. The next day it may not work, but their experience now is, “Oh this works,” and so they’re not prepared to take over and so their ability to kind of save it and monitor it decays with time. So you know in America, somebody dies in a car accident about 1.15 times per 100 million miles. That’s like 10,000 years of an average person’s driving. So, let’s say the technology is pretty good but not that good. You know, someone dies once every 50 million miles. We’re going to have twice as many accidents and fatalities on the roads on average, but for any one individual they could go a lifetime, many lifetimes before they ever see that.” – https://www.recode.net/2017/9/8/16278566/transcript-self-driving-car-engineer-chris-urmson-recode-decode https://www.recode.net/2017/9/8/16278566/transcript-self-dri... or Ford Motor Co. executive vice president Raj Nair – “Ford Motor Co. executive vice president Raj Nair says you get to 90 percent automation pretty quickly once you understand the technology you need. “It takes a lot, lot longer to get to 96 or 97,” he says. “You have a curve, and those last few percentage points are really difficult.” Almost every time auto executives talk about the promise of self-driving cars, they cite the National Highway Traffic Safety Administration statistic that shows human error is the “critical reason” for all but 6 percent of car crashes. But that’s kind of misleading, says Nair. “If you look at it in terms of fatal accidents and miles driven, humans are actually very reliable machines. We need to create an even more reliable machine.” – https://www.consumerreports.org/autonomous-driving/self-driving-cars-driving-into-the-future/ https://www.consumerreports.org/autonomous-driving/self-driv... or prof. Raj Rajkumar head of Carnegie Mellon University’s leading self-driving laboratory. – “if you do the mileage statistics, one fatality happens every 80 million miles. That is unfortunately of course, but that is a tremendously high bar for automatically vehicle to meet.” min.19.30 of this podcast interview – http://www.slate.com/articles/podcasts/if_then/2018/05/self_driving_cars_are_not_yet_as_safe_as_human_drivers_says_carnegie_mellon.html http://www.slate.com/articles/podcasts/if_then/2018/05/self_... or Elon Musk himself - "On average, a person is killed in a traffic accident in the United States once every 100 million miles. Elon Musk says Tesla’s Autopilot is half as likely to be involved in a collision as a human driver. That would suggest that somewhere around the 200 million mile mark someone will die as a result of an automobile driven by a machine." - https://www.teslarati.com/autonomous-cars-make-us-worse-drivers/ https://www.teslarati.com/autonomous-cars-make-us-worse-driv... What you are using is a fallacy, emotional statement done by self driving cars developers and enthusiasts in order to make people think by adopting this technology they will be part of a bigger better future, by doing essentially nothing."
  - grecy 6y ago
    What I'm saying is that self-driving cars don't have to be perfect, they just have to be better than human drivers. Once there are less deaths with self-driving cars that human-driven cars (however you want to measure that) then it makes sense to use it.
    - uniqueid 6y ago
      > they just have to be better than human drivers. That's where the difference in our outlook is: The way I see it, people can stomach the existence of ridiculous accidents (ridiculous in the sense that the details are "absurd", not just "very grim") when they can blame a human. I don't think many people will tolerate even occasional "self-driving" accidents if the circumstance show a human would have recognized the danger. For a self-driving car to be fit-for-purpose, it can't run over kittens at low speeds on an empty road - even if they're very small kittens that don't exceed some motion-sensing threshold. That requires common-sense reasoning.
      - grecy 6y ago
        I think you're absolutely right there will be push-back if the self driving cars crash is stupid ways, but again, it's all about the numbers. Planes still fall out of the sky for stupid reasons (maintenance errors, mostly) yet the numbers say they're very, very VERY safe, so people use them. Once the facts show self driving cars are safer than human driving cars (even if they sometimes crash in stupid ways), there will be no holding back.