tyrion

That's a good article. But all he's really saying is that we don't have solid evidence that self driving cars are safer than human driven cars. That is true, but to put it in perspective, there was a time when we didn't have solid evidence that machines were better than humans at keeping banking records.

RidesaJapanese

If by 300 million you mean the US population, that logic has a flaw. How many of those 300 million are regular cyclists?

RidesaJapanese

Maybe the bigger problem is this just isn't a bicycle friendly country, and with few viable alternatives to the automobile there are a lot of them on the roads, most of which are not designed with cyclists in mind, or poorly so. If this country had a modern rail net like Europe or Asia, I'd scrap my beater car in a heartbeat. I live in a region where small stagnating towns all have a railroad history, but no railroad any more. Maybe I'm too idealistic, or visionary, to imagine riding my bike to the local light rail, and then being somewhere in a few hours that would take all day driving. That's just being un-American, implying that we're not exceptional and do it better than anywhere else.

I-Like-To-Bike

Put in perspective, your response sounds like others posted on this list - when there is no solid evidence to support hyped PR about almost ready for prime time self driving cars, just chant Smartphone! or some other similar non sequitor.

tyrion

Have I chanted "smartphone"? Is what I "sound like" really something to base your argument on? Can you point out any non sequiturs I've used on this subject?

I-Like-To-Bike

"there was a time when we didn't have solid evidence that machines were better than humans at keeping banking records." Just as fallacious and irrelevant as the Smartphone and similar non sequitors referenced by other posters on this list.

Maelochs

Actually .... completely accurate. Before computers were developed, no one knew that data storage, data manipulation and data access would be literally millions of times faster using computers. No one knew ...

Fallacious? Only if you don’t know what “fallacious” means.

He points out quite plainly that we do not have any data on whether AI cars will be safer than human-driven cars. (We do have Plenty of data proving human-driven cars are unsafe—in fact people here whine about that almost as much they whine about the AI cars which don’t yet exist.)

He is implying that in time we Will know that AI cars are safer ... and for rational people, it does seem that AI cars will be better at driving than drunk drivers, texting drivers, road-ragers deliberately hitting other cars, drivers spacing out in traffic and driving into the car ahead ... and likely better at coping with emergency situations, when all too many drivers just scream, lock their arms, mash the brake, and close their eyes.

But he never States that ... he sort of implies that. So line about bank records is in no way “fallacious” at all, even slightly. Nor is it irrelevant, as I have explained (You do know what “irrelevant means, I hope?)

However ... Your post about Now, That is fallacious. Nobody here has said AI cars are “almost ready for prime time.” No One ... but you. That is a flat lie you invented because you couldn’t win the argument using honest means ... and you couldn’t let it go.

And “non-sequitur”? Only if you don’t know what “non sequitur” means.

A "non sequitur" is a phrase which does not in any way follow logically from the previous phrases (and if you want to play an intellectual on the internet, spellcheck is your friend ... Google is your lord protector.)

And no one here has chanted “smart phone” .... but even that would not be a “non sequitur” because everyone here can see that you are refering to the many posters who mention how much of the tech used in smartphones will be easily adaptable (or has in all likelihood been adapted) to AI cars.

People who have been reading the thread can understand the reference ... so that is not a non sequitur. Pointing out that smartphone tech is similar, follows logically, and your mockery of it follow by your own twisted logic.

No “non-sequiturs” or even ”non-sequitors” here, friend.

So ... in two brief posts you call an accurate statement fallacious, tell lies, and prove that you do not know what “non-sequitur” means ... and cast doubt of your grasp of the meaning of “irrelevant.” Yeah ... that certainly raises your standing around here. A Lot more people will take you seriously now.

So again ... when you want to preach from the intellectual high ground, you might want to actually reach the intellectual high ground before you open your mouth.

Or maybe not, based on history ....

So Anyway ...

At this point, all anyone is saying is “I think AI cars can be made to work,” or “I don’t think AI cars can be made to work.”

Considering that several dozen AI cars have logged hundreds of thousands of miles so far with no fatalities and the vehicles are still in their infancy, it would seem that the issues are not insurmountable, but the final evidence is not in.

So .. it is all opinion.

And once each of us have stated our opinions ... all that is left is what we have here: “You’re wrong, I’m right.” “No, you’re wrong, and I’m right,” ad nauseam.

But thanks for the entertainment.

Not much progress has been made in this thread and I think no more ever will be.

Rollfast

All of these responses and their complex theories and wishful thinking and it still boils down to a distinctly possible dystopian reality where roving anarchist gangs surround the cars and rob and harm the passengers.


As I've been saying for like almost five years now. I want to run at full speed before I die from the explosion.


That's that. And yes, I'm sure it occurred to the people trying to flee the false incoming missile in Hawaii that it's a hydrogen bomb, you are going to die anyway.


There is the old line about I want to die like grandpa, behind the wheel with three screaming passengers.


Sorry to be your Debbie Downer.

Rollfast

The simple human brain is still far more capable than the computers we create *or they would start killing us like the constructs of Westworld.


Keeping the organic mind in it's place as the marvel of operation that it is can always be of great help. You are operating under the notion that we understand all aspects of our thinking, which is patently untrue, and misleading. All brains operate under the template of their learning environment as well as emotional needs and complex feelings and then there are the defects as no mind is standard or perfect but some reflection of it's creation and care.


As an example, I can express my thoughts well yet make a lot of typing errors which I must correct, including clarifying the context or how I used a word, possibly incorrectly or in a malapropism.


PS There is a difference between a 'brain' and an 'autonomic nervous system', if I am to take the road of reasoning I gather you use. While it is true that the action of steering and controlling acceleration and braking is fairly straightforward, the judgments that facilitate these reflex behaviors certainly are NOT.

genec

And yet that wonderful human brain still manages to become easily distracted, and somehow rationalizes driving at excessive speeds for the road conditions as well as engaging in irrational behavior under the influence of alcohol... all of which leads to about 40,000 deaths a year.

I suspect you are expecting far too much from the average driver, and over estimating the skills required to drive.

Maelochs

BS. It is Differently abled.

Computers can store and access data and can compare and calculate millions of time faster than you. Computers beat the greatest human chess masters. If your brain is really all that, turn off your computer and keep this conversation going.

Cannot fly a B2 bomber without computers---because they can sense attitude changes and respond precisely, much, much faster than a human pilot could.

Why?

In any case, one doesn't need a computer to commit murder, individual or en masse ... so that is a big "fail."

I am sure that the families of the 30,000 to 40,000 people killed in auto accident each year since about 1960 agree with. (Ninety-Nine point Nine percent driver error ... Hurray, humans! using cars to kill is obviously something our brains have mastered.)
I don't see where anybody is saying anything even vaguely like this. Yes, while computers can be programmed to write music, most music nowadays is written by humans using computers. Computers cannot generally paint as well as humans---but if programmed, they can paint Better than humans in many ways (for instance, for a photorealist painting, a computer will be more realistic.) A computer can apply an algorithm to interpret a scene is any of a number of styles and produce a painting in that genre (think Photoshop filters.)

So far the human brain is better at pure creativity ... but only the tiniest portion of the human race creates anything but misery.

Here is the point you are Not mentioning: The stuff a computer does millions or billions of times better than a human are the things needed to drive a car.

It goes back to the B2-bomber example. Precise sensing, mass processing and precise response is all the realm of the computer (and attendant computer-controlled servos.)

The reason we are not already all driving AI cars is that along with gathering all the sensors, feeding all the data to the processor, and writing all the code which makes all the "if/then" decisions .... is that companies cannot make cars as big as trucks and want to make money at it. If the military wanted AI cars, we'd have them.

Yup. But neither is it impossible to do. The reason several dozen AI cars have Already logged hundreds of thousands of miles is because Most of the base code is written.

With every new scenario, appropriate code is written. Then eventually all that code is debugged and resolved .... as it has been with airplanes, which (according to pilots I have communicated with) can pretty much do the whole flight on autopilot almost all the time.
Which exactly why computers should eventually be able to do the job better.

Computer will not get angry at other drivers, stay mad over fights with significant others, get angry about the kids yelling in the back seat or the guy talking on the radio, or just life in general. Computers won't get drunk, won't get stoned, and won't have any trouble handling texts and phone calls while driving.

Computers won't space out, get sleepy, get so into singing along with the radio they forget they are driving, won't deliberately do dangerous things to express their frustration with life itself. Computers won't play "race driver." Computers will not deliberately endanger people of different nationalities, social classes, displaying certain bumper stickers, or people using different transport modes, simply out of dislike for those categories.

Computers will not reject factual information which they do not "like" because they do not "like" or "dislike." Computers do not distort information or inject extraneous data into calculations in an attempt to produce an solution which is "pleasing." Who do you think does that?

Computers do have computer-specific issues. And in general, those can be addressed through hardware and software.

Did it not occur to anyone that almost every word we type to one another gets to its audience because of giant rockets mostly designed by, mostly constructed, by, mostly launched by, and along trajectories mostly calculated by, computers, which deployed, according to computer programs, immense communications satellites, and which, despite micro-meteorites and cosmic radiation and all that, still manage to handle immense streams of data and direct them to the right electronic receivers. When something goes wrong, it is almost always because a power pole fell ---often after being hit by a drunk in a car.

AI cars might not take over tomorrow ... but people who think it cannot be done ... well, there is that low-quality, emotion-limited, protein computer showing its weaknesses.

squirtdad

you know we all keep forgetting about how self driving bikes would help.....with the Netherlands leading

Ninety5rpm

+1 on your whole post, but especially the above.

And slight quibble with this:

I'm not that cynical. The military does want AI cars, I'm sure. Taking vehicles (including tanks!) into combat without risking lives of military personnel? Of course they want that!

We're getting AI cars as fast as reasonably possible. The best minds in the world are working on it, feverishly. It's the biggest technological race since getting into space and to the moon, the PC and the web - probably bigger.

MinnMan

Not sure I want to jump into this maelstrom, but I thought that this recent Christina Bonnington piece might add some perspective

https://slate.com/technology/2018/02...he-answer.html

Ninety5rpm

+1

MinnMan

Never mind. Just saw that that article has its own thread.

As you were.

I-Like-To-Bike

I'll make it easy for you; no word salad necessary to respond.
The Smartphone reference arguments used on this list were NOT that similar technology may also be incorporated in self driving cars therefore somehow greasing the skids for future success of a related product.

The "Smartphone argument" as used on this list are that citing the successful and profitable (at least for Apple) deployment of smartphone technology (or any other irrelevant product/unrelated product) on the market obviates any doubt, skepticism or argument about the inevitable successful deployment of self driving vehicles, AND/OR obviates any doubt, skepticism or argument about the assumed givens that such vehicles will be cheaper to operate (or provide equivalent transportation utility to passengers at less cost), AND/OR obviates any doubt, skepticism or argument that the deployment of self driving vehicles will reduce injury and fatality rates while providing equivalent transportation utility to the public.

Citing a "completely accurate" (and completely irrelevant) statement that "solid evidence that machines were better than humans at keeping banking records" fits the fallacious BF Smartphone argument category when used to discuss the evidence or lack of such whether self driving cars will be safer and reduce injury/fatality rates.

Citing the success of Smartphones and record keeping machines and other equally irrelevant arguments, whether accurate or not, (buggy whips are SO old fashioned) to support an opinion or assumption about the future success of self driving vehicles ARE NOT logical, ARE fallacious and most definitely ARE non sequitur arguments.

Thanks for the spell checking commentary though.

I-Like-To-Bike

Here is a new related article that may be of interest:
https://spectrum.ieee.org/cars-that-...getting-better

True believers may wish to ignore any discussion about human intervention frequency during the testing of so-called driver less vehicles.

Ninety5rpm

This statement is key:Initial deployments will be within confined areas where all routes are thoroughly vetted to not be problematic for autonomous vehicle operation. These regions will grow and eventually overlap and connect, ultimately covering the entire country.

As each new area is mapped out for the first time, some disengagement, human intervention and machine learning is expected. Of course.

noisebeam

They do exceptionally well detecting vehicular cyclists who signal their intentions and sometimes not. I've been training 'em well.

I-Like-To-Bike

Yes that is possible, especially if deployment within a confined area means REALLY confined to areas that present NO problems for their operation , and "eventually" and "ultimately" are considered open-ended guesses about future developments.

Me? I think these statements are "key":

Rollfast

As soon as I fix mine...PS I just got some Laserdiscs in the mail today as well, but I'm not selling any watches...

genec

Still using buggy whips too, no doubt.

prathmann

No, but I was just playing around with some of my slide rules yesterday.

fietsbob

that Truck that ran down people on the NYC bike path could have been a remotely operated Drone ...