Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I work with AI and Machine Learning every day. The goal is to avoid the crash with any object regardless of what the object is. Because if you hit a vehicle versus a person you may actually cause more deaths and injuries because the vehicle you hit plows into a crowd. Unintended consequences are very difficult to predict.

Indeed, but I do hope the AI can take a good guess at what it's seeing, as, even the most perfect system, may need, at some point, to decide the best of two worse scenarios.
And how solid something is. Tree trunk, Brick Wall, Softer Hedge.

If a child runs into the street from between parked cars with no time to stop (perhaps wet road) Veering into the hedge to avoid the child would seem to be the better option.
 
Honestly, AI is a dangerous road to go down, read Darknet by Matthew Mater for an extreme example of what can go wrong.

We keep asking "Can we make self driving cars?" and we fail to ask "Should we make self driving cars?"

I keep hearing this but look at the current state of Siri, Alexia, and every other algorithm based "AI" and just laugh.
 
Last edited:
I agree especially if it's driving me around

100% agree.

IMHO there are many possible problems in autonomous driving.

Just to name a few:

- can it be used for targeted crashing of specific person or specific persons
- if the car crashes, who is the priority for saving, person in car or higher number of persons
- who is responsible for potential crashes, company who made AI, company who sold it, owner
- trolley problem
 
  • Like
Reactions: lowendlinux
Yeah I'm totally confident that a self-driving car will be able to properly determine right-of-way... Especially in a place like a complex parking lot...

I work with AI and Machine Learning every day. The goal is to avoid the crash with any object regardless of what the object is. Because if you hit a vehicle versus a person you may actually cause more deaths and injuries because the vehicle you hit plows into a crowd. Unintended consequences are very difficult to predict.
Except the law also states that you don't stop in the middle of moving traffic for certain things. Like it's a good idea for a car to slam on the brakes just because it thinks something is about to happen...

I have zero faith in this kind of tech. I still get app crashes on my iPhone, even on Apple's own apps.
 
This is good for everyone. Apple sharing this research instead of keeping it proprietary shows they are more interested in getting to a sustainable future faster than retaining whatever competitive business advantage this would give them.

They're more interested in keeping their researchers. Top ones would leave if they couldn't publish.

In this case, the paper is only about a theoretical way of doing something a bit easier. It's not something necessary.

No doubt they still won't be able to publish things that would provide substantial business advantages.
 
Last edited:
IMHO also only SAE level 5 is the only safe version of autonomous car driving for regular people.

I think SAE level 3 and lever 4 will be / are problematic and would/could also increase the number of accidents, because automation systems will decrease the ability of drivers due to over reliance on those systems.
 
One step at a time.
And who asked for self driving cars? Since humans are quite good at causing terrible accidents on the road that costs the life of millions each year, self driving car seems the logical solution. Who wouldn't want to minimise the road accidents?

I feel like reducing this statistic starts by getting rid of infotainment systems in vehicles - like CarPlay. Then next by designing better roads. And maybe reconsidering expanding public transportation (in the US) outside of major metros.

There's a lot of low-hanging fruit if your goal is to reduce auto-related deaths. All three of those things I listed above would be immediately available to everyone, regardless of their socio-economic circumstances. Self-driving cars are (IMO) hot-air, a solution looking for a problem.
 
If "relax" is the main advantage of self driving cars, how about using public transport?
I drive 5km to the nearest bus stop, take the bus 20 minutes to the train station, then the train to the city where I work. From the central station, it's a 10 minute walk. All in all, it takes about 2 hours.
Yes, driving would cut that time in half, but hey, I can already enjoy the benefit of self-driving cars: relax.
For me, the difference between self-driving and driven by someone else is none.
 
  • Like
Reactions: PizzaBoxStyle
"I can find the pedestrian with SEVEN sensors."
"I can find the pedestrian with SIX sensors."​
"I can find the pedestrian with FOUR sensors."
"I can find the pedestrian with ONE sensor."​
"Find that pedestrian!"

Haven't we heard this tune before? :)
 
Except the law also states that you don't stop in the middle of moving traffic for certain things. Like it's a good idea for a car to slam on the brakes just because it thinks something is about to happen...

I have zero faith in this kind of tech. I still get app crashes on my iPhone, even on Apple's own apps.

I don't understand your response. Would you do as a driver in the situation of something impeding your path? Would you just continue to drive and hit the object. Or would you slow down and wait for a transient object to pass (child, dog, car pulling out in front of you) or drive around it?
 
I honestly never want a self driving car. I have little faith in computers being able to reliably do something so dynamic. I hope theres never any forced self-driving cars only laws put into place.

Even if self-driving cars were barely half as good at humans at dealing with all the stupid driving mistakes made by humans, if we got rid of all the human-driven cars on the freeways, they would immediately be orders of magnitude safer than humans, because they wouldn't have to deal with any real surprises, and would have much, much better reaction time when dealing with the car ahead of them slowing down or speeding up. This would mean a much faster average speed on freeways, because:
  • They can maintain a much shorter distance to the next car safely.
  • Their faster reaction time would nearly eliminate the accordion action that contributes so heavily to traffic backups.
  • Their multiple cameras allow them to merge more accurately, causing fewer delays.
  • They don't fall asleep at the wheel.
  • They don't exceed the maximum safe speed for the road.
  • They don't aggressively, angrily try to pass other vehicles and cause worse backups behind them when they cut people off (and sometimes get hit).
So I want a self-driving car yesterday. Nearly all car accidents are directly attributable to human error, and I have little faith in humans ever being able to reliably do something so dynamic with only two eyes that can only point in one direction at a time. They're hopelessly crippled compared with machines. I hope there is a law within the next six months requiring all cars to have cameras and sensors built-in to make future self-driving upgrades possible, and I hope that human-driven cars are restricted to under-35-MPH city streets within the next ten years, if not sooner.
 
  • Like
Reactions: jerryk
Indeed, but I do hope the AI can take a good guess at what it's seeing, as, even the most perfect system, may need, at some point, to decide the best of two worse scenarios.
And how solid something is. Tree trunk, Brick Wall, Softer Hedge.

If a child runs into the street from between parked cars with no time to stop (perhaps wet road) Veering into the hedge to avoid the child would seem to be the better option.

Or perhaps it would not be a better option. The best thing about an autonomous car would be it may have detected that child very early and slowed down to avoid the conflict.

Car vision/sensor system are far superior to humans. Seeing in darker conditions than we can, able to detect and classify objects at much farther distances, and in 360 degrees at once. Then this information can be used to determine how to avoid a conflict.
[doublepost=1511381483][/doublepost]
IMHO also only SAE level 5 is the only safe version of autonomous car driving for regular people.

I think SAE level 3 and lever 4 will be / are problematic and would/could also increase the number of accidents, because automation systems will decrease the ability of drivers due to over reliance on those systems.

That is what Google/Waymo believes. The problem with less than level 4 is the human is the backup. And since the human may not be paying attention, the human is not in sync with the environment and will not what do when the system says "here, you take over".

It is interesting times. Cadillac/GM will be shipping their super cruise in 2018 model year cars which can handle freeway driving. Uber is going self drive with the 20,000 Volvos they are buying. Their OTTO division is putting self-drive into semis.
 
Last edited:
I honestly never want a self driving car. I have little faith in computers being able to reliably do something so dynamic. I hope theres never any forced self-driving cars only laws put into place.

It will likely be decades before self driving cars are in the majority. But I bet soon as that happens people will pass laws to keep human driven car off major streets and urban areas.

People will be terrified at the idea of human driven cars. And they should be. Today a human driven car is the NUMBER ONE CAUSE OF DEATH for people under the age of about 40.

I think it is reasonable to bad something that is the #1 killer.

Self driving cars don't need to be perfect, just good enough that they are say the #2 cause of death would be a huge improvement.. World wide about 1,000,000 people die in car accidents. IN the US the number is 20,000. Think of the World Trade Center attack on 9/11. Cars kill as many people as 10 of those attacks happening every year. Of course they will ban human driven cars just as soon as such a ban is practical.
 
  • Like
Reactions: ErikGrim
I don't understand your response. Would you do as a driver in the situation of something impeding your path? Would you just continue to drive and hit the object. Or would you slow down and wait for a transient object to pass (child, dog, car pulling out in front of you) or drive around it?
Well for example, if you're driving down the freeway and a coyote runs out in front of your car, you're required to hit the coyote if necessary. If you hit the brakes or swerve and cause an accident then you are at fault.
 
Well for example, if you're driving down the freeway and a coyote runs out in front of your car, you're required to hit the coyote if necessary. If you hit the brakes or swerve and cause an accident then you are at fault.

Never heard of a law that requires you to hit something rather than swerving or braking around it. And if someone hits you from behind when you hit your brakes, they are usually found at fault for following too closely. There is a reason the "3 second guideline" exists.

But the good thing about a self driving cars is that their systems are focused 100% on object conflict detection. And have better detection capabilities that we will ever have. Some systems they can tell the difference between a piece of paper and a ball at 200 yards. And if they do have to swerve they know which way is less likely to cause a conflict since they are constantly checking for conflicts to the left, right, in front ,and behind, and have no blind spots.
 
Last edited:
Imagine how many aholes would dedicate huge amounts of their time and resources into trying to hack those cars and cause chaos. And there‘s no reason to believe that not at least some of them would succeed.
[doublepost=1511391062][/doublepost]
Never heard of a law that requires you to hit something rather than swerving or braking around it..

It‘s not a law, but insurances might require you to do so. AFAIK they do so here in Germany.
 
It‘s not a law, but insurances might require you to do so. AFAIK they do so here in Germany.
That sounds crazy. Would if it is a deer? Those often kill people because they go right through the windshield into the face of the driver and front passenger. And what if the person is riding a motorcycle?
 
That sounds crazy. Would if it is a deer? Those often kill people because they go right through the windshield into the face of the driver and front passenger. And what if the person is riding a motorcycle?

Of course it‘s crazy to hit a deer instead of evading into a field at the side of the road. But unless the deer keeps standing there until the police or whatever arrive, it‘s going to be hard to make the insurance believe you that there was one in the first place.
[doublepost=1511392011][/doublepost]
No, it doesn't mean that at all.

OK, I admit it was worded stupidly provocative.
But please explain:
Why should we expect the AI to work better than the one powering Siri, and how are cloud features working with no reception?
 
Never heard of a law that requires you to hit something rather than swerving or braking around it. And if someone hits you from behind when you hit your brakes, they are usually found at fault for following too closely. There is a reason the "3 second guideline" exists.

But the good thing about a self driving cars is that their systems are focused 100% on object conflict detection. And have better detection capabilities that we will ever have. Some systems they can tell the difference between a piece of paper and a ball at 200 yards. And if they do have to swerve they know which way is less likely to cause a conflict since they are constantly checking for conflicts to the left, right, in front ,and behind, and have no blind spots.
I can promise you that fault isn't so cut and dry.
And when the self driving car is involved in the accident, how much more uncertainty will that add?
[doublepost=1511392330][/doublepost]
That sounds crazy. Would if it is a deer? Those often kill people because they go right through the windshield into the face of the driver and front passenger. And what if the person is riding a motorcycle?
If it is big enough it would create an actual emergency. The problem is in the "what constitutes an emergency" that you had to brake/swerve.
Sorry mods for double-posting. edit: did it just add it in automatically? cool
 
Why are they making these? Who has ever asked for self driving cars?

We wanted FLYING cars, I think they got this mixed up somewhere.

Have you driven on the road lately? Self driving cars can reduce acccidents, driver fatigue and keep traffic moving. Today a light turns green and some cars crawl out people are on phones and don’t go at all. People run red lights and cause accidents, cars merge into other cars people fall asleep. Now imagine a future we’re the light turns green and every car moves at the same time and pace. Imagine all cars stopping at a red light. Imagine on the interstate we’re all cars are going the same speed and paced at the same distance. Imagine a car driving itself in stop and go traffic. Now imagine when the make flying cars with self driving tech and you don’t need a pilots license.
 
  • Like
Reactions: jerryk
Honestly, AI is a dangerous road to go down, read Darknet by Matthew Mater for an extreme example of what can go wrong.

We keep asking "Can we make self driving cars?" and we fail to ask "Should we make self driving cars?"

What happens if we allow human to drive their own cars? How could anything be worse then what we have now? Cars are the number one cause of death if you are under about 40. Guns kill far fewer people than cars

Anyone who worries abut the safety of the self driven car has to compare it to the safety of a human driven car and importantly have numbers to back up the claims.

An example of what could go wrong does not matter. What matters is how frequently accidents happen. We don't have enough data on this yet.
 
  • Like
Reactions: jerryk
Sharing also means hopefully others can contribute to it to improve over Apple as well
 
This is good for everyone. Apple sharing this research instead of keeping it proprietary shows they are more interested in getting to a sustainable future faster than retaining whatever competitive business advantage this would give them.
Academic types weren’t all that willing to go work at Apple if they couldn’t publish.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.