Driverless Hype Collides with Merciless Reality • Hacker News Board

Re: Driverless Hype Collides with Merciless Reality

#221

Post by xyhopguy » Thu, Sep 13, 2018, 9:33 PM UTC

01100011 wrote: post #201

pkulak wrote: post #170

Earlier quoted context omitted.

No one thinks that you just hook up ML to a camera and you're done. Literally everyone working on this only uses ML for identification. When that Model X crashed into a highway divider, it was because the lines on the road were very faint and there was a fork in the line that actually wasn't the _real_ lane line. The car followed that by mistake, but only because of the misidentification of the lane line. How are mor…

Well, my optical sensors(i.e. eyes) provide me with enough information to discriminate a cat from a bag, so I suppose with enough processing we should be able to do the same with a control system. Cars have an added advantage that their vision systems could work in lower light and non-optical wavelengths. I don't see any fundamental issues with getting self-driving cars working. It may not happen in 5 years, but it w…

The tough part about this line of reasoning is you can't tease apart the discriminatory power of human knowledge from human vision. How good does our expert system need to be? How good do our sensors need to be? I'm not so sure this is an easy to question to resolve

Re: Driverless Hype Collides with Merciless Reality

#222

Post by amalcon » Thu, Sep 13, 2018, 9:36 PM UTC

leesec wrote: post #178

tree_of_item wrote: post #158

Earlier quoted context omitted.

How would we get accurate information on the number of lives Tesla's "autopilot" has saved?

Humans die on average every 80 million miles driven ( not sure the metric for highway miles only ), but prove that Tesla can beat that metric and you can assume lives are being saved.

In addition to the confounding variables others have mentioned, I think the most important factor is when autopilot is enabled. If you don't (or can't) use autopilot at night, in inclement weather, or in construction zones, then you'll get into more accidents when not using autopilot -- even if it's a significantly worse driver than you are.

Re: Driverless Hype Collides with Merciless Reality

#223

Post by Fricken » Thu, Sep 13, 2018, 9:36 PM UTC

linuxftw wrote: post #217

Fricken wrote: post #212

Earlier quoted context omitted.

The two key aspects of an autonomous OS are perception and planning. Being able to "see" well enough during a flurry is a perception problem. Knowing whether to maneuver out of it or call for help after spinning out is a planning problem. There are many planning problems left to address and master before autonomous vehicles are ready for widespread adoption. Some planning problems are unique to specific intersections…

Can they perceive where on the road they should be located when the road is totally covered in snow and you can't see the pavement, let alone the line markets, possibly only tire tracks, if that?

Yes. An autonomous vehicle with lidar and maps can locate itself using any landmark, or series of landmarks it can see, it isn't limited to lane markings.

Re: Driverless Hype Collides with Merciless Reality

#224

Post by dunpeal » Thu, Sep 13, 2018, 9:36 PM UTC

justinpaulson wrote: post #154

fmpwizard wrote: post #83

Earlier quoted context omitted.

If we look at humans as a counter, sure, apply all the stats you'd like, but I don't think we should do that when it comes to self driving cars. if we simplify the driving risks, we have: 1. driver kills him/herself 2. driver kills person outside his/her car * Humans and computers can make mistakes on both. But, my wife and I can teach our kid everything we have learned over the years, to prevent 1, and try to reduce…

Everyone thinks they are good drivers, yet hundreds of thousands of people are dying every year because of these "good drivers." Not sure what makes you think that teaching your kid what you know will make them a better driver than software that will continue to be iterated on and improved based on millions and millions of miles of data (more data than you and your wife could ever pass along to your child). Also not…

Exactly. People used to think computers will never beat humans in chess, clinging to all sorts of romantic notions such as that inherently human understanding was required for high-level play.

And now my phone can beat the world champion.

Re: Driverless Hype Collides with Merciless Reality

#225

Post by baybal2 » Thu, Sep 13, 2018, 9:38 PM UTC

icebraining wrote: post #203

baybal2 wrote: post #100

Earlier quoted context omitted.

My point is that the popular image of flying cars with retractable wings and wheels turning into jet engines is ludicrous, and that yes, light personal airplanes and helicopters are there. And the same parallel for driverless car, except for even when well executed it is kinda of arguable usefulness/value.

arguable usefulness/value Are regular cars useful? A self-driving car is just as useful for someone who can't drive: minors, many seniors, many people with disabilities - or people who just don't have a driving licenses, like me.

For them, a 30km/h crawler with overattentive collision avoidance system may well be useful. This is not much different to how public transport drivers have to drive.

I expect them facing same issues like trams and trolleybuses - terribly bad experience in traffic.

You only see risky people aggressively trying to overtake and scratching their paint, while leaving the driver no opportunity to move an inch.

Re: Driverless Hype Collides with Merciless Reality

#226

Post by beat » Thu, Sep 13, 2018, 9:39 PM UTC

Alupis wrote: post #136

yannyu wrote: post #85

Earlier quoted context omitted.

Or maybe we hold driverless cars up to an unfair standard. Today, 3000 people die in the US every single day because of car accidents. Many, many more are injured or disabled. Why would we expect driverless cars to reduce that to 0? Shouldn't it be good enough if they reduce it to 50%? 25%?

> Today, 3000 people die in the US every single day because of car accidents Whoa... that's over 1 million people annually... you probably mean globally? Depending on the year, it's somewhere between 30,000 - 40,000 annually[1][2][3], or in the worst-case, about 109 people daily in the US. When you examine those numbers further, you'll see about 10,000-18,000 (depending on the year) are sadly still alcohol related[4]…

Hey, you know what would fix the drunk driving problem?

Autonomous cars.

Re: Driverless Hype Collides with Merciless Reality

#227

Post by SketchySeaBeast » Thu, Sep 13, 2018, 9:40 PM UTC

icebraining wrote: post #203

baybal2 wrote: post #100

Earlier quoted context omitted.

My point is that the popular image of flying cars with retractable wings and wheels turning into jet engines is ludicrous, and that yes, light personal airplanes and helicopters are there. And the same parallel for driverless car, except for even when well executed it is kinda of arguable usefulness/value.

arguable usefulness/value Are regular cars useful? A self-driving car is just as useful for someone who can't drive: minors, many seniors, many people with disabilities - or people who just don't have a driving licenses, like me.

For those of us who do drive it would be great to get that commute time back.

Re: Driverless Hype Collides with Merciless Reality

#228

Post by telchar » Thu, Sep 13, 2018, 9:41 PM UTC

dragontamer wrote: post #194

beat wrote: post #182

Earlier quoted context omitted.

The 100% attention thing I why I have tremendous faith that computers will be out-driving humans in short order. Even really good human drivers have a limited perspective and focus. And even the best human drivers are subject to emotional state.

Yeah, humans suck at attention. But until we can program computers to tell me what is a foreground moving object, what is a shadow, and what is the background... you literally can't solve the "Should the car apply the brakes right now" problem. https://arxiv.org/pdf/1801.02225.pdf Just go through this recent paper I found on DDG.gg and see the amount of effort it takes to parse foreground / background data that a sel…

Humans suck at these classification tasks things too, though, if you set your expectations high. Is it a cat or a bag, or a weird shadow? Doesn't matter, I ran it over before I could figure it out and react. Or maybe it turned out to be a boulder? Now I'm one of 35,000 traffic fatalities this year in this country.

Computer vision has been improving rapidly in the last 10 years, I think it's too soon to rule out the viability of a camera-based solution entirely. Though I do hope improved lidar technology can improve on humans.

Re: Driverless Hype Collides with Merciless Reality

#229

Post by linuxftw » Thu, Sep 13, 2018, 9:41 PM UTC

Fricken wrote: post #223

linuxftw wrote: post #217

Earlier quoted context omitted.

Can they perceive where on the road they should be located when the road is totally covered in snow and you can't see the pavement, let alone the line markets, possibly only tire tracks, if that?

Yes. An autonomous vehicle with lidar and maps can locate itself using any landmark, or series of landmarks it can see, it isn't limited to lane markings.

And if there are no landmarks, such as you're on a highway or rural area, not in a city?

Re: Driverless Hype Collides with Merciless Reality

#230

Post by Fricken » Thu, Sep 13, 2018, 9:46 PM UTC

linuxftw wrote: post #229

Fricken wrote: post #223

Earlier quoted context omitted.

Yes. An autonomous vehicle with lidar and maps can locate itself using any landmark, or series of landmarks it can see, it isn't limited to lane markings.

And if there are no landmarks, such as you're on a highway or rural area, not in a city?

Not a ditch, or a barbwire fence, or a telephone pole or a tree? I guess that may be tricky.