Elon Musk: Smartest Guy in The World or Complete Clown?

**riser4** · 07-02-2016, 11:42 AM

I'm just glad that he only removed his genes from the pool and not anyone else's. Benny might be a doosh on teh TGRz, however, I still don't want one of these idiots mowing down some cyclists. Fuck autonomous driving.

sb · 07-02-2016, 12:27 PM

Originally Posted by iceman

That stuff only makes sense if you're the one playing the game. And if I was playing a game there would be way more hookers and blow. So it doesn't make sense.

This

**rod9301** · 07-03-2016, 12:09 AM

Driverless cars is not like releasing software, where you incrementally improve the software until it is acceptable.

You have to have an automated car perfect from the beginning.

Which is why Google is testing on millions of miles until they release a product.

**abraham** · 07-03-2016, 12:27 AM

Originally Posted by rod9301

Driverless cars is not like releasing software, where you incrementally improve the software until it is acceptable.

You have to have an automated car perfect from the beginning.

Which is why Google is testing on millions of miles until they release a product.

Well Tesla has been treating it that way. Kind of. And as a result they have more data from more miles than Google. I think by orders of magnitude. But I agree that it is tricky, probably the best similar scenario is the introduction of elevators? But then people in Germany have achieved historical protection for paternoster elevators, so that may not be a good example.

I do know of at least one person who left google's autonomous car program because of ethical concerns about writing code that could kill. And I know a few others who were uncertain about staying on the project. But that type of uncertainty was visible in several projects that were pushing the limits of technology and dragging us along with it, so I can't judge them harshly.

**shroom** · 07-03-2016, 12:32 AM

my point was that any such software will be iterated upon [probably multiple times daily] after it's released. obviously, you can reduce risk by doing as much testing as possible without humans involved.

**abraham** · 07-03-2016, 12:43 AM

Originally Posted by shroom

my point was that any such software will be iterated upon [probably multiple times daily] after it's released. obviously, you can reduce risk by doing as much testing as possible without humans involved.

You sound like some sort of software nerd.

And like you know what you're talking about.

I think my biggest problem is with how the autopilot was marketed, but then it's still very much a startup and hype is important. We do some projects that tangentially involve nuclear waste and there are all sorts of verification checks and tests, but they are only as good as the specs and sometimes the people writing them don't think about checking trivial things. And that's in a very controlled environment, driving is much harder for sure

**shroom** · 07-03-2016, 12:51 AM

hah, I ain't no pro, but I've seen enough to know that any software will have bugs, even though tesla hires some of the brightest minds around. my car has auto windows but not much else.

and you hit the nail on the head with marketing. biznass wants product in consumer faces, so you see a lot of testing in production; MVP and all. I would hope an industry like this would avoid that type of approach as much as possible.

**Bobby Stainless** · 07-03-2016, 03:52 AM

Regardless of cars being autonomous or not, accidents will still happen.

Cars will have to weigh human life, and in most cases will kill the "driver", as opposed to a group of pedestrians.

The continual saving of our society, is tedious.

The loss of jobs for truck drivers, cabbies, Uber, bus drivers, etc, will be inevitable. Pointing to the bigger issue the world has, especially China, of lack of work.

**old goat** · 07-03-2016, 08:18 AM

When they start testing autonomous cars in Truckee in the winter I'll believe that they might work.
In the meantime, I'll just be glad the Tesla idiot took himself off the road permanently without killing anyone else.
On second thought--maybe it's a good thing. People make turns on red, run red lights, change lanes, make lefts across traffic and pull out into traffic, counting on oncoming cars to hit the brakes. Maybe if they think the oncoming car might be a Tesla whose "driver" is having sex in the back seat (do Teslas have back seats?) they might wait until traffic clears.

**Benny Profane** · 07-03-2016, 08:49 AM

Originally Posted by old goat

When they start testing autonomous cars in Truckee in the winter I'll believe that they might work.
In the meantime, I'll just be glad the Tesla idiot took himself off the road permanently without killing anyone else.
On second thought--maybe it's a good thing. People make turns on red, run red lights, change lanes, make lefts across traffic and pull out into traffic, counting on oncoming cars to hit the brakes. Maybe if they think the oncoming car might be a Tesla whose "driver" is having sex in the back seat (do Teslas have back seats?) they might wait until traffic clears.

Around here, running stop signs is common, and I'm starting to see left turns on red more and more.

**Stu Gotz** · 07-03-2016, 06:04 PM

Releasing production misses on July 3rd? As my first boss used to say , "F*cking Clown Car."

**frorider** · 07-06-2016, 09:49 AM

http://electrek.co/2016/04/20/elon-m...topilot-video/

**muted** · 07-06-2016, 11:18 AM

38,300 people were killed last year on US roads. Tesla has only killed one guy this year, who was masturbating to a Harry Potter video. Not quite time to wring hands about autopilot, yet.

**MagnificentUnicorn** · 07-06-2016, 12:05 PM

I was talking to one of the surgeons I work with last night about the "adaptive cruise control" on his Tesla. He said it's pretty good going straight, on the freeway and on two lane roads as long as the fog and center lines are really well marked. He also said it's great cruising over to Seattle once you're on the freeway but that he wouldn't feel comfortable not paying attention. Interesting to get a real world perspective.

**old goat** · 07-06-2016, 12:15 PM

Originally Posted by AaronWright

I was talking to one of the surgeons I work with last night about the "adaptive cruise control" on his Tesla. He said it's pretty good going straight, on the freeway and on two lane roads as long as the fog and center lines are really well marked. He also said it's great cruising over to Seattle once you're on the freeway but that he wouldn't feel comfortable not paying attention. Interesting to get a real world perspective.

around here the lane and fog lines are gone by mid winter, and are frequently covered by snow and ice before that.
the more I think about it the more I think the engineers working on this stuff are grossly underestimating the intelligence and ability to deal with novel situations of even the dumbest human brain and grossly overestimating their ability to duplicate it, let alone improve on it. I believe autonomous cars will do better than people at routine driving situations--like the ones you describe--but introduce new failure modes when confronted with the infinite variety of novel situations human drivers handle without thinking about it. To truly succeed will require not only greatly improved autonomous vehicles but massive infrastructure spending to eliminate as much as possible the novel situations-for example sensors or radiomagnetic reflectors in all roads. IMO the investment would be better spent on a comprehensive upgrade of mass transit.

**MagnificentUnicorn** · 07-06-2016, 12:24 PM

Originally Posted by old goat

IMO the investment would be better spent on a comprehensive upgrade of mass transit.

This is what I keep telling people when the subject of autonomous vehicles comes up. The reason I hear most for wanting autonomous vehicles is so that they don't have to pay attention and they would be able to read, socialize and watch movies or browse the internet. Seems like those are activities for the bus and train.

**Skidog** · 07-06-2016, 12:34 PM

Originally Posted by old goat

around here the lane and fog lines are gone by mid winter, and are frequently covered by snow and ice before that.
the more I think about it the more I think the engineers working on this stuff are grossly underestimating the intelligence and ability to deal with novel situations of even the dumbest human brain and grossly overestimating their ability to duplicate it, let alone improve on it. I believe autonomous cars will do better than people at routine driving situations--like the ones you describe--but introduce new failure modes when confronted with the infinite variety of novel situations human drivers handle without thinking about it. To truly succeed will require not only greatly improved autonomous vehicles but massive infrastructure spending to eliminate as much as possible the novel situations-for example sensors or radiomagnetic reflectors in all roads. IMO the investment would be better spent on a comprehensive upgrade of mass transit.

My thoughts exactly. Of course a computer like the IBM chess playing computer can kick a humans ass at chess...there are only so many move combinations possible. They can all be programmed or searched and compared against. With driving there are far too many variable that a human can handle much faster than a computer.

**old goat** · 07-06-2016, 06:31 PM

why don't they just train chimpanzees to drive? they're smarter than computers. and you can pay them peanuts (pay them peanuts, get it?).

**BmillsSkier** · 07-06-2016, 06:54 PM

Originally Posted by old goat

why don't they just train chimpanzees to drive? they're smarter than computers. and you can pay them peanuts (pay them peanuts, get it?).

But then who would do all the writing in Hollywood?

**babybear** · 07-06-2016, 06:55 PM

Originally Posted by BmillsSkier

But then who would do all the writing in Hollywood?

Amd manage all of our money?

**The Suit** · 07-06-2016, 07:25 PM

Originally Posted by Skidog

Of course a computer like the IBM chess playing computer can kick a humans ass at chess...there are only so many move combinations possible. They can all be programmed or searched and compared against.

Wrong. Number of possible combinations is greater than the number of atoms in the observable universe.

**Benny Profane** · 07-06-2016, 07:33 PM

Originally Posted by old goat

why don't they just train chimpanzees to drive? they're smarter than computers. and you can pay them peanuts (pay them peanuts, get it?).

No, sorry, were way beyond chimps in computer development. Watson kicked everyone's ass on Jeapordy, and that was what, four or five years ago? Watson's probably plotting world domination by now.

I agree about mass transit, but, that is so 180 degrees from what those libertarian geeks out in that valley are imaging for the world. They probably get all grossed out having to fly with other humans in coach, or riding the company bus every day out of the Mission.

Back to Musk, I heard a cruel comparison of him to Bernie Ebbers the other day, and maybe that pundit had a point. Not that he's going to jail some day, but, the house of cards is looking more and more ready to fall, despite the hype. He better have spent a lot of money on Hillary's campaign. As we know, she's easily bought.

**Bromontana** · 07-06-2016, 07:41 PM

Originally Posted by The Suit

Wrong. Number of possible combinations is greater than the number of atoms in the observable universe.

This is 2016, not even Central Banks use math.

**iceman** · 07-06-2016, 08:08 PM

Originally Posted by The Suit

Wrong. Number of possible combinations is greater than the number of atoms in the observable universe.

If it had to consider every ppossible move that would be relevant but of course only an infinitesimal percentage of the total possible moves in a chess game are available as options at any time, even with the opening move. The computer's just crunching possibilities for chess. Now, Go...