With this category of blowhard (Musk, Trump, Kanye West) I think it’s a self-reinforcing cycle of saying something rhetorically and then starting to believe your own bullshit.
I’m glad someone already posted this observation. Also common in this category are hordes of supporters who will explain away whatever their hero said based on what paints them in the best light at the moment.
Oh come now. I respect musk and trump. But I don't deny both say ridiculous shit that makes you wonder why they said it. Most supporters I know of either are willing to acknowledge their faults and shortcomings. I can still respect someone who does apparently stupid shit if in the end they produce what others couldn't or wouldn't.
They are still humans. If they like to talk themselves up to do things then so what?
Musk and West are certainly in a different level of achievement in non-marketing activities than Trump (and he is in a different level than them in marketing achievements). However along this parameter of puffery I see nearly identical psychologies — hyperdefensiveness, an obsession with “haters,” and a complete exaggeration of social proof.
To add context to the above, looks like SpaceX's Falcon Heavy launched 2018-02-06 [1] after Elon Musk's 2011 claim [2]:
> “Falcon Heavy will arrive at our Vandenberg, California, launch complex by the end of next year, with liftoff to follow soon thereafter. First launch from our Cape Canaveral launch complex is planned for late 2013 or 2014.”
I mean, I suppose this doesn't really have all that much to do with automated driving. But yes, I do support much greater liability for the people or entities that operate motor vehicles.
I think you should focus on civil penalties, not criminal. Criminal penalties for bad driving would not work in America, politically. Besides, higher insurance minimums help to make the cost of operating dangerous machinery around other people be born by the people who are choosing to operate that machinery. much greater insurance requirements and greater civil penalties will raise the cost of driving to include the cost of the injuries caused by driving
(right now, where I am, the insurance minimum is like a month or two of software engineer wages... so obviously, your insurance bill doesn't cover the likely cost if you actually injure or kill a person)
Of course, this would make driving a lot more expensive (well, it would shift the cost away from the people who are hit to the people who are choosing to drive these dangerous machines) and realistically would only work in urban areas where there is decent public transit. (Or maybe you could do something with geofencing insurance; you are a lot more likely to hit a person in an urban area, so maybe insurance in low-population areas would be cheaper, if you never drove in areas with a lot of other people?)
If anything, I think self-driving cars would make this slightly better, as the only sane way to insure them is for the company that created and maintained the self-driving system insure it, and that means there are deeper pockets behind the thing; deep enough to pay for the loss of a human life, or to repair a human.
I'm guessing this is a tongue in the cheek response to this: [1]. I think that clever people always run a risk of trying to be "too clever" and self driving cars is the prime example of our generation.
A simpler solution would be to install electronic wiring on the roads themselves to guide the cars and I daresay we would have self-driving cars within a month of installation. Now, I am not saying we should do this (South Africa, I'm scolding you for thuggery and vandalism) but it does support my argument about "clever people" having collective detrimental group think.
> A simpler solution would be to install electronic wiring on the roads themselves to guide the cars and I daresay we would have self-driving cars within a month of installation.
Many billions of dollars have gone into developing self-driving car tech. But carving up our entire road network to install signaling equipment? That would cost trillions.
Whoever said that we need to do the entire road network?
A few geofenced areas, like a dozen-blocks in Manhattan, would probably be profitable. Start with the densest cities, 90% of the profits with only 10% of the effort. Write an Uber-like app so that people would know whether or not their starting-location + destination are within the fence, and you're basically done.
Why? Self-driving cars can already do those dozen blocks in Manhattan. And a lot more than that. The wait currently is to ensure every reasonable edge case is handled, and to understand the degree of safety.
And how does a wire in the road help avoiding pedestrians, bad human drivers, construction, etc? You still have to build intelligence into the car.
Idk how you envision this system working for millions of vehicles. In all sorts of weather. What about pedestrians? What robust connection would allow this wire to do its job?
Honestly I think developing the technology for cars to drive themselves as is currently being developed is a much more elegant and feasible solution than what you're proposing.
And trams don't do what we need done. They go in small straight lines, mostly in low speed areas, still require a driver, cost a ton, and as I said - don't avoid running over humans.
Since these have been around for decades (I think I saw a video from the 1960s about a future world of self-driving cars with such sort of system), it's likely not so easy to do.
> In 1957, a full size system was successfully demonstrated by RCA Labs and the State of Nebraska on a 400-foot strip of public highway at the intersection of U.S. Route 77 and Nebraska Highway 2, then just outside Lincoln, Nebraska. A series of experimental detector circuits buried in the pavement were a series of lights along the edge of the road. The detector circuits were able to send impulses to guide the car and determine the presence and velocity of any metallic vehicle on its surface. A previous test installation of the system in September 1954 along U.S. Route 73 and U.S. Route 75 in Cass County, Nebraska was utilized as an experimental traffic counter. It was developed in collaboration with General Motors, who paired two standard models with equipment consisting of special radio receivers and audible and visual warning devices that were able to simulate automatic steering, accelerating and brake control.[27][28][29]
I guess he sets impossible deadlines not for the public but for his employees. He thinks he has a very limited time and has a lot to achieve. It is fine until someone loses money because of this news.
It can't be very pleasant for his employees to have a CEO that throws around those numbers publicly though. It seems like at some point down the management hierarchy there must be a disconnect where people above the line are on board with the official timeline and people below the line are quietly planning a more reasonable timeline. Like "Elon thinks that we'll launch the Heavy in two years but that's obviously impossible, so don't burn yourself out trying to get it done." This sounds like a recipe for organizational dysfunction as people above the line are frustrated by the apparent lack of progress and wastefully provide the wrong level of resources given the actual status of the project.
TLDR; Fun exchange between Elon Musk and George Hotz. Later put a bet he can do better than Mobileye. Musk declined citing difference between demoware vs production. Musk hopes to have level 4 autonomy in two years.
30 comments
[ 1.9 ms ] story [ 103 ms ] threadAnd even better it is partially work and kill a few ...
They are still humans. If they like to talk themselves up to do things then so what?
Which could easily have the title:
"Elon Musk Says SpaceX Will Launch Falcon Heavy In Two Years (2011)"
Sure, don't believe his dates, but he may just deliver it.
> “Falcon Heavy will arrive at our Vandenberg, California, launch complex by the end of next year, with liftoff to follow soon thereafter. First launch from our Cape Canaveral launch complex is planned for late 2013 or 2014.”
[1]: https://en.wikipedia.org/wiki/Falcon_Heavy_test_flight
[2]: https://www.spacex.com/press/2012/12/19/spacex-announces-lau...
Auto vs assist. Force vs option. Blind-Trust vs not-be-Lab-rat
Ok if anyone kill all programmer and the company executive is subject to murder charge. How about that?
I think you should focus on civil penalties, not criminal. Criminal penalties for bad driving would not work in America, politically. Besides, higher insurance minimums help to make the cost of operating dangerous machinery around other people be born by the people who are choosing to operate that machinery. much greater insurance requirements and greater civil penalties will raise the cost of driving to include the cost of the injuries caused by driving
(right now, where I am, the insurance minimum is like a month or two of software engineer wages... so obviously, your insurance bill doesn't cover the likely cost if you actually injure or kill a person)
Of course, this would make driving a lot more expensive (well, it would shift the cost away from the people who are hit to the people who are choosing to drive these dangerous machines) and realistically would only work in urban areas where there is decent public transit. (Or maybe you could do something with geofencing insurance; you are a lot more likely to hit a person in an urban area, so maybe insurance in low-population areas would be cheaper, if you never drove in areas with a lot of other people?)
If anything, I think self-driving cars would make this slightly better, as the only sane way to insure them is for the company that created and maintained the self-driving system insure it, and that means there are deeper pockets behind the thing; deep enough to pay for the loss of a human life, or to repair a human.
A simpler solution would be to install electronic wiring on the roads themselves to guide the cars and I daresay we would have self-driving cars within a month of installation. Now, I am not saying we should do this (South Africa, I'm scolding you for thuggery and vandalism) but it does support my argument about "clever people" having collective detrimental group think.
[1] https://arstechnica.com/cars/2019/04/elon-musk-says-tesla-is...
Many billions of dollars have gone into developing self-driving car tech. But carving up our entire road network to install signaling equipment? That would cost trillions.
A few geofenced areas, like a dozen-blocks in Manhattan, would probably be profitable. Start with the densest cities, 90% of the profits with only 10% of the effort. Write an Uber-like app so that people would know whether or not their starting-location + destination are within the fence, and you're basically done.
And how does a wire in the road help avoiding pedestrians, bad human drivers, construction, etc? You still have to build intelligence into the car.
Honestly I think developing the technology for cars to drive themselves as is currently being developed is a much more elegant and feasible solution than what you're proposing.
Trams already work.
"Look Ma, No Hands! Automated Bus Steers Itself" - https://www.wired.com/2008/09/look-ma-no-hand/
"Volvo Thinks Magnetic Roads Will Guide Tomorrow's Autonomous Cars" - https://www.wired.com/2014/03/volvo-magnets-autonomous/
Guided vehicles on factory floors have been around for a while, see https://patents.google.com/?q=guided&q=vehicle&oq=guided+veh... .
Since these have been around for decades (I think I saw a video from the 1960s about a future world of self-driving cars with such sort of system), it's likely not so easy to do.
> In 1957, a full size system was successfully demonstrated by RCA Labs and the State of Nebraska on a 400-foot strip of public highway at the intersection of U.S. Route 77 and Nebraska Highway 2, then just outside Lincoln, Nebraska. A series of experimental detector circuits buried in the pavement were a series of lights along the edge of the road. The detector circuits were able to send impulses to guide the car and determine the presence and velocity of any metallic vehicle on its surface. A previous test installation of the system in September 1954 along U.S. Route 73 and U.S. Route 75 in Cass County, Nebraska was utilized as an experimental traffic counter. It was developed in collaboration with General Motors, who paired two standard models with equipment consisting of special radio receivers and audible and visual warning devices that were able to simulate automatic steering, accelerating and brake control.[27][28][29]