955 replies to this topic

[marcobjj]

Joshua D. Brown, of Canton, Ohio, died in the accident May 7 in Williston, Florida, when his car’s cameras failed to distinguish the white side of a turning tractor-trailer from a brightly lit sky and didn’t automatically activate its brakes, according to government records obtained Thursday.

Frank Baressi, 62, the driver of the truck and owner of Okemah Express LLC, said the Tesla driver was “playing Harry Potter on the TV screen” at the time of the crash and driving so quickly that “he went so fast through my trailer I didn’t see him.”

http://www.breitbart...m_medium=social

[mag1]

I did not understand that story when I first read it.

 

The narrative seems to be suggesting that the car's autopilot was to blame, even while the driver appears to

have floored the accelerator. I would tend to think that all of these automated driving systems are specifically 

programmed to strictly follow all posted speeds.

 

It also seemed very strange to realize that this category of accidents might be entirely preventable if all cars were

required to have onboard GPS beacons. In that way, every object on the roads could be identified and visible to

the automated navigation systems. Perhaps cyclists and even people (dogs?) might want to have such GPS beacons

in order that they would also be "seen" by the autonavs and correct risk assessments be applied in each instance.

 

It is difficult for me to accept this incident as an example of technology being less capable than people: it actually tends to show

the opposite.

Edited by mag1, 05 July 2016 - 09:46 PM.

[mag1]

I have read some additional information on this accident.

 

Overriding the autonav's adherence to posted speeds appears to be quite possible.

Online reports are suggesting that the car involved might have been traveling at 85 MPH or higher.

The highway in question was non-controlled access.

 

Some ideas that spring to mind: I wonder whether a supercomputer could run through almost every possible configuration

of vehicles to detect whether other such blind spots might exist with the technology. I would also be interested to know 

whether cars could be equipped with some sort of simple lighting system that could shine (perhaps a distinctive symbol,

a corporate symbol ?) onto any objects in the path of the vehicle and have an onboard computer determine whether a

solid stationary or slowly moving object were ahead (Might even consider putting a simple reflective decal or other

observable design on all vehicles which could be recognized by the onboard autonav system.). For whatever reason

the system (radar) used by the car in this collision did not recognize that an object was ahead.

 

This accident also highlights the overwhelming advantage of computer learning: the ability to learn from such an incident 

and then redesign the technology so that this specific accident scenario will never be repeated. Unfortunately

humans do not have such a networked intelligence. Clearly humans have been making similar driving errors resulting

in an enormous number of road accidents for many decades.

   

Edited by mag1, 06 July 2016 - 12:08 AM.

[resting]

Move the sensors up.

Or add whiskers.

[mag1]

Upon further reading about this story I find it deeply troubling that even after the entire roof had been sheared off

and the driver was no longer responsive the car continued to drive on without any attempt at braking. Shouldn't there

be a sensor which would indicate that a major accident had occurred and the car could then bring itself to a controlled

stop? Online reports have noted that if the car had stayed on the highway that it is quite possible that it could have

traveled for many miles until remote control would have attempted braking. Not having an automatic braking protocol

in such circumstances is a severe design flaw. 

Edited by mag1, 06 July 2016 - 02:33 AM.

[mag1]

I have an idea.

 

Why not create an online database for the times at which these 18 wheelers make these turns into oncoming traffic?

 

The drivers of these rigs likely know weeks in advance the exact routes they will be taking and they typically

know these routes very very well. They probably could provide times of their turns to the minute. 

 

The rig operators could post the times and exact locations of their turns online and then other travelers

could access this information. A virtual right of way could be given to the truckers for perhaps a 1 minute window.

A half mile exclusion zone could be created in the approaching traffic.

 

Underriding accidents involving tractor trailers and sedans are among the most traumatic of all vehicle accidents.

Creating a database could be one way to reduce these accidents.

 

If others are interested in this idea: Start posting!

This could go viral.

 

Truckers would seem to have a considerable amount to gain by posting this information.

They would be acting constructively to solve a problem that has not been adequately addressed in decades.

 

This is not a complex idea, though it should help make our roads safer.

Let's see some support for this!

 

After the people have voted with their posts we can call for some adult supervision and get the programming

in motion.

 

 

 

 

 

 

 

 

 

 

 

[mag1]

I think this is a great idea!

Could someone provide some feedback?

 

These tractor trailer accidents are often extremely severe.

The cars are sometimes nearly completely destroyed.

 

They are intending on creating an intelligent highway system, though this could take years and years. 

There is no reason to wait years.

 

My idea would take advantage of a low level human based communication and interaction network to leverage information that is now knowable.

It seems like a highly unwise idea to have no level of organization on our existing highways.

Consider how much safer our highways would be if every time an 18 wheeler turned into an oncoming lane of traffic that all the cars gave a half mile of clearance

and all were under the speed limit.

 

My idea could start being used immediately!

There is no reason that this has to be sent to a committee.

This could be a true grassroots effort to make our highways safer.

 

 

Comments Please!

 

 

Edited by mag1, 07 July 2016 - 03:44 AM.

[niner]

I don't think that truck traffic is sufficiently consistent that one could know in advance when they would cross a road with enough time resolution to be useful.  The tragic Tesla Autopilot incident was a case of Pilot Error.  The car was being operated outside of its limited zone of safe use, (i.e. closed-access freeways), the driver was speeding and watching Harry Potter instead of the road.  Truck-car collisions of this sort are not all that common.  There are many other ways that driving could be made safer that would be more effective. 

[mag1]

I think this is a doable idea.

 

It is simply a virtual stop light.

The trucker could pull into the divided median, push a button on his cab, and then

after perhaps a minute or two wait would be given virtual right of way.

Drivers that ignored such a right of way would have committed a driving offense and could be fined and

demerit points assessed.

 

The problem with the current system is that there is no particular reason other than

being nice for drivers to allow the trucker through the intersection. This creates the

circumstances for the recent fatality. If this idea had been in place the fatality would

not have occurred. The onboard software in the car would have received notice that

right of way had been granted to the trucker and the car could have alerted the driver

of the car and started a controlled deceleration. 

 

 

"The United States Department of Transportation estimates that over 500,000 truck accidents occur every year."

"Nearly 5,000 people are killed in truck accidents every year." (that is in the US)

 

http://www.truckinfo...cking/stats.htm

 

The true cost of not moving ahead with my idea is actually much higher.

From what I have read the intention is to handicap driving technology by perhaps a factor of 2 (or even 10).

With such a handicap, 20,000 people per year will die every year by allowing humans to control their vehicles and not technology.

The latest results from Tesla indicate that the factor of 2 handicap threshold might already have been crossed.

 

"Estimates of 41,000 to 45,000 traffic deaths occur every year within the U.S.."

 

Any ideas such as the one that I have suggested that incrementally improve safety would allow us to cross the line and gain the reward of 20,000

fewer accident fatalities per year. So the gain would be greater than only the trucking fatalities themselves. Of course, this would also allow sooner

roll out on a global scale. The world experienced 1.25 million road fatalities in 2013.

 

http://www.who.int/g...aths_number/en/

 

The experts in autonomous car navigation have noted that the hold back in perfecting the technology has a lot to do with the outlier incidents that

are nearly unpredictable. Simply driving on a clear day on a well marked highway where the drivers are traveling the speed limit should be a technologically

achievable goal. Filling in all of the outliers could take a considerable period of time. However, by finding ways in which we could leverage human

interaction in the driving experience, then we could roll out the new technology sooner which would allow for a much greater of driving experiences to be

acquired leading to an acceleration of the development of the technology.  

 

 

 

Edited by mag1, 07 July 2016 - 06:50 PM.

[niner]

What you're proposing is vehicle to vehicle communication on a grand scale.  V2V has been under discussion and study for well over a decade, and will cost many billions even if it's gradually rolled out over 15 years.  It's not a cost-effective or quick solution to the current problem.  One thing that would be cheap and easy is to mandate daytime running lights on all new cars, and possibly to retrofit them.  They are required in Canada and the EU, but NHTSA cites a study that claims they don't save lives.  I don't buy the "study", and neither does much of the free world.  I think this might have something to do with lobbying efforts by the US auto industry.  Another way to reduce accidents would be to equip traffic signals with sensors that could see oncoming traffic, along with a modicum of intelligence to reduce the number of unnecessarily-stopped vehicles and light changes against oncoming vehicles.  This would work immediately for all vehicles, not just those that were intelligent, and could be profitably rolled out in a piecemeal fashion to the most dangerous intersections first.  It would reduce accidents, but also save fuel and time, increase the carrying capacity of intersections, and reduce stress on drivers.

 

Tesla doesn't claim their system is a true autonomous vehicle, and I suspect after the present debacle they will make it harder to drive too fast with autopilot engaged, and should take steps to disable entertainment systems visible to the driver if that hasn't already been done.  There are probably a lot of other things they could or should do to make the system safer.

[sthira]

Stupid people spoil good things. This dumbshit and his Harry Potter.

Meanwhile, has anyone mentioned SF's possible opening of Robo Burger?

Robo-Burger Joint to Open in San Francisco

http://futurism.com/...&post_type=link

Watch out Burger King, a robot-run burger restaurant is opening in the South of Market area of San Francisco.

According to a Craigslist job posting looking for a restaurant manager to lead the robo-dining experience, “this location will feature the world-premiere of our proprietary and remarkable new advances in technology that enable the automatic creation of impossibly delicious burgers at prices everyone can afford.”

Momentum Machines

The robo-restaurant manager will still take customers’ orders, schedule shifts, and occasionally take out the trash and tidy up. As an added bonus, managers can learn skills like software troubleshooting, market research, and product development research.

The burger-making robots, invented by Momentum Machines in 2012, can customize every aspect of the burger, from thickness and cook time to condiments. Some critics worry that the robots may put a lot of people out of work. But Momentum Machines says jobs will be created by providing opportunities in restaurant management and technology development.

[mag1]

What I find particularly interesting is that no great investment would be needed to move my idea forward.

 

The truck driver could auto dial for his virtual Right of Way with perhaps a few simple clicks on his cell phone or tablet.

Perhaps the privilege would cost $1 - $2 (or even better provided gratis by interested parties).

 

The cell phone could give the GPS coordinates, the driver might indicate that the intended direction of

motion would be for example west (probably best to have some way to double check direction perhaps using the directional

orientation of the cab) and the intended time to exercise this right of way (perhaps say from now until one minute

from now, voided after it has been exercised).

 

Notification of the granting of the Right of Way could then be sent to cell or tablets in the proximity. The whole process

might only take a few seconds. An app on the device could simply give out an audio alert such as "Virtual Stop Light

10 seconds ahead". Cars might be given a 30 second notification period, so that those already close to the truck could

pass through as if it were a yellow light.

 

None of this would involve any great expense. The $100,000 vehicle involved in the recent accident

might only need a software upgrade to enable such a feature. This could all be done by a private entity, government 

might not even be required to be involved. This is simply a private contract agreed by these parties. By observing these

contracts the roads would be safer for everyone.

 

One of the big problems with technology is people will often wait for the super duper implementation of a technology when

even a minimalist, low cost version would offer substantial benefits and could be completed years ahead of the ultimate version.

 

How many million lives will be lost while we wait for the self-driving car 2.0?

 

 

 

 

 

 

Edited by mag1, 07 July 2016 - 09:57 PM.

[niner]

What I find particularly interesting is that no great investment would be needed to move my idea forward.

 

The truck driver could auto dial for his virtual Right of Way with perhaps a few simple clicks on his cell phone or tablet.

Perhaps the privilege would cost $1 - $2 (or even better provided gratis by interested parties).

 

The cell phone could give the GPS coordinates, the driver might indicate that the intended direction of

motion would be for example west (probably best to have some way to double check direction perhaps using the directional

orientation of the cab) and the intended time to exercise this right of way (perhaps say from now until one minute

from now, voided after it has been exercised).

 

Notification of the granting of the Right of Way could then be sent to cell or tablets in the proximity. The whole process

might only take a few seconds. An app on the device could simply give out an audio alert such as "Virtual Stop Light

10 seconds ahead". Cars might be given a 30 second notification period, so that those already close to the truck could

pass through as if it were a yellow light.

 

None of this would involve any great expense. The $100,000 vehicle involved in the recent accident

might only need a software upgrade to enable such a feature. This could all be done by a private entity, government 

might not even be required to be involved. This is simply a private contract agreed by these parties. By observing these

contracts the roads would be safer for everyone.

 

As usual, I'll be the grouch who says this isn't going to work:

 

1)  Truck driver already knows how to cross the road, doesn't want to fuck with a phone or tablet then wait 30 seconds when he could have been on his way a minute ago.

 

2)  Driver on the road already has the right of way relative to the truck that is crossing.  Driver has no interest in stopping in order to cede his right of way and risk getting rear-ended by the car or truck behind him that doesn't have the right app and isn't expecting them to stop in a random location for no apparent reason.

 

3)  This system only works when EVERYONE has a device with the correct app, that device is charged and on, and the driver knows how to use it.  Otherwise, the virtual red light is useless because half the drivers won't know it's there and will blow right through it.
 

[mag1]

It is always good to think through the objections before moving something ahead.

 

I have seen these truckers caught in the median trying to cross over a few lanes of traffic and they were stuck for longer than a minute.

It must be so frustrating for them. At some point they must just figure I am going for it. In the modern world no one is going to cede

Right of Way. If I were a trucker given the choice of perhaps a minute or two wait and having to cross without any warning, then I would definitely

choose the Virtual Right of Way option. The trucker would never be obliged to move into traffic even with Right of Way if it wasn't thought safe.

In the present accident almost the worst imaginable scenario was selected by the trucker because an almost random car was chosen to 

cross against.

 

 

When you look at the accident scene it is not even clear whether the trucker could contain his full rig in the median section. Having to

do a standing start acceleration into 2 lanes of high speed approaching traffic with an 18 wheeler is not something that I would ever want to try.

If you gave me a tablet with a Virtual Right of Way app, then I probably would.

 

The proper implementation of the idea would not require any actual stopping. In one embodiment it would simply entail slowing down enough to give 

the trucker space to make the cross. It might require motorists to slow to perhaps 40 MPH in a controlled manner. Interestingly if this had happened in 

the present scenario, the autonav would have had no problem recognizing that cars were slowing ahead and would have alerted the driver.

 

The underlying problem in these circumstances is that approaching traffic always has Right of Way. If no one cedes it then the trucker is going to become

increasingly frustrated. It is best not to get someone with an 18 wheeler mad.

 

Creating a Virtual Right of Way transfers Right of Way to the Trucker. Motorists that did not recognize this Right would be endangering all other drivers on the

road, financial and other sanctions could then be applied to them. As it is now motorists who never cede Right of Way and create these dangerous circumstances 

never face consequences.

 

It does not seem unreasonable to suggest that such an app (probably free) could become mandatory. There are all sorts of laws concerning the equipment that is required 

to operate a motor vehicle. In the modern world, requiring an app with some sort of device ( tablet etc.) would probably entail a financial cost to few motorists.

Driving is not a constitutionally protected right. Even if there were those who did not have the app, seeing an 18 wheeler in the median

would be understood by motorists to mean that a crossing would soon occur.

 

 

It is also notable that this entire discussion is now being driven by a fairly small number of drivers who have spent upwards of $100,000 for their vehicle and are receiving frequent

software updates for their "cruise control system". The driver involved in the current accident was a technology expert. Those who might drive in a manner that placed them

at higher than average risk of these crashes would likely be highly motivated to upgrade to the latest possible gear. Those other drivers who habitually drove in the slow lane and

tried to maintain a speed 5-10 MPH under the limit would likely be unaffected by any changes related to Virtual Right of Way.

Edited by mag1, 08 July 2016 - 01:49 AM.

[mag1]

Much of this self driving technology is being rolled out by stealth.

Reports have indicated that some of these semi-autonomous trucks were tested on US roads.

 

The article stated that Daimler was going to commit 530 billion euro by 2020 for automated transport.

Was that a typo?

That is an extraordinary amount of money!

 

 

http://www.roadandtr...-rig-convoy-on/

Edited by mag1, 09 July 2016 - 01:10 AM.

[PWAIN]

Look like the Chinese are going to be the first to test Crispr in humans:

https://www.theguard...ng-trial-humans

[albedo]

You might be interested:

 

Where machines could replace humans—and where they can’t (yet)

http://www.mckinsey....e-they-cant-yet

[PWAIN]

Albedo, interesting article. Not sure if they are considering the use of the latest deep learning. These systems change the landscape in my view and will result in far more work and work types being automatable. Do they consider level 4 self drive? That is not really a tech available today but seems almost certain to be with us within the next 5 to 10 years if not sooner.

[albedo]

Albedo, interesting article. Not sure if they are considering the use of the latest deep learning. These systems change the landscape in my view and will result in far more work and work types being automatable. Do they consider level 4 self drive? That is not really a tech available today but seems almost certain to be with us within the next 5 to 10 years if not sooner.

I agree Pwain and I am not sure they do, they should. They look to base the analysis on currently demonstrated technologies available commercially or through research project but maybe not that advanced: "We define “currently demonstrated technologies” as those that have already exhibited the level of performance and reliability needed to automate 1 or more of the 18 capabilities involved in carrying out work activities. In some cases, that level of performance has been demonstrated through commercially available products, in others through research projects." Deep learning will likely increase both the no. of capabilities and the reliability.

[mag1]

So glad that at least Germany has got it all together.

We certainly need to get ahead of bad technologies such as drone copter-plane delivery before

we experience a perpetual eclipse of the sun.

 

The buggy drones make so much more sense and appear to now be rapidly approaching.  

 

http://www.thelocal....ure-of-delivery

[Elus]

'We're just rentals': Uber drivers ask where they fit in a self-driving future

 

"With Uber set to deploy autonomous cars in Pittsburgh, some drivers for the company wonder if they’ve been expendable all along"

[Jason Burns]

One way I look at it, is that robots & computers don't have to replace 100% of human work in a job to have a big impact on employment.  For example, in an area where 100,000 people work, if automation replaces 40% of human work over several decades, it will mean 40,000 workers no longer needed in that area.

 

While the country will still be getting the benefit of the former 100,000 output, while having 40,000 surplus workers to deploy elsewhere.  The problem for most of the population, is that the new areas of opportunity that are opening up require high levels of intelligence.  

 

How I see Western governments responding to this so far, is by creating legions of government jobs that are of questionable utility.  An administration for a vast government agency could expand from 40,000 workers, to 80,000 workers over 20 years and hardly anyone would notice.  

 

 

[mag1]

Jason, I think that merely the projected future implications of technological change is having a big impact. Globally total fertility rates plunged in most industrialized nations decades ago in anticipation of the world that we are now rapidly approaching. Some of the early adopters of a low or no fertility lifestyle were some of the more intelligent in our communities. The implications of this loss of genetic intellectual potential will play out in the decades to come. The idea of Singularity by itself is already generating a large and increasing social force.

 

The geopolitical implications of decades and decades of advanced industralized societies living below replacement levels are all too obvious to see.

The Rise and Fall of Human Civilization would not be a completely erroneous characterization. Societies that are unable to reproduce sufficiently become

highly vulnerable to large scale migrations into their nations and the strains that this can cause. From the broadest historical persceptive it is not clear whether

this decline is salvagable. Perhaps the 1980s will continue to be seen nostigically as a golden era for many years. It is always possible that everything will work

out fine, though this is by no means a certainty.

 

It is only recently that we have seen the beginning of what might turn out to be irreversible population declines in great nations such as Russia and Japan. Without large scale immigration most other developed nations would be in the same circumstance.

 

The idea that those displaced by technology can somehow find make work employment in government is somewhat questionable. Governmnet services

could quite possibly be some of the most vulnerable to a profound rethinking. Technology has already been developed and is rolling out that could greatly

disrupt the model of government monopoly services.

 

Furthermore, part of the plunge in fertility rates is perhaps due to the idea that parents to be have assessed the employment prospects of their potential offspring and decided that these prospects were not sufficiently promising enough to have children. The idea that in the future people could just go to make believe jobs or be paid to exist is likely not enough justification for some to bring a life into this world. Might not some question their self worth, if it were obvious that they simply were unneeded in the labor economy? How might human rights be impacted if people are not actually being productive?

 

Such logic has important implications for humanity: as human populations shrink in developed nations the totality of human potential declines with it.

If 1 person in 10 billion is an Einstein, then perhaps we will not have that one Einstein.                

Edited by mag1, 21 September 2016 - 10:35 PM.

[marcobjj]

Mass immigration is the biggest scam of the 21st century. If you wonder why Egypt and Persian that were at one time the top civilizations on the planet have become third world hellholes. Well they suffered the same demographic transition that is taking place in Western Civilization now. They were overrun by low IQ arabs who imposed their primitive religion and culture upon everyone else. Europe will suffer the same fate if current trends are allowed to continue.

 

All of these aging countries, Japan, Western Europe, are much better off subsidizing their own families then importing people from the middle east who are taking welfare anyway. If we wish for intelligent life to continue on in this planet, then the prime time civilizations must be preserved at all costs. 

[mag1]

It is a great concern to me that we are now within sight of a radical restructuring of our economy with robotic transport.

Reports now suggest that there is broad agreement that this will happen and will roll out over the next 5-15 years.

 

I have proposed several times on this thread that robotic transport would remove the central spoke of our economy.

There would be no great reason then for the entire distribution of goods to not occur from mega boxes somewhere near to

airports / shipping ports. For those not particularly interested in in store shopping there would be no great need to

lug oneself to a store.

 

This is only one of the profound changes that are now clearly foreseeable.

I can only wonder at what point there might be yet another fertility collapse in response to such changes.    

Edited by mag1, 23 September 2016 - 11:34 PM.

[mag1]

Peepull! Peepull!

 

We have been sleeping on our job as humanity`s advanced warning system for the approaching Singularity break out moment.

Once the fuse is lit there is no turning back.

 

Perhaps this has now happened.

 

"It begins."

Is that as ominous as I think it is?

 

http://www.scienceal...-its-own-memory

 

A somewhat obscure publication though,

http://www.nature.co...ature20101.html

 

 

 

 

Edited by mag1, 16 October 2016 - 06:49 PM.

[william7]

 I was reading the passages below tonight and just had to share it with you folks on this site. It's from The Revolution of Hope: Toward A Humanized Technology (1968), by Erich Fromm.

 

One symptom of the attraction of the merely mechanical is the growing popularity, among some scientists and the public, of the idea that it will be possible to construct computers which are no different from man in thinking, feeling, or any other aspect of functioning. The main problem, it seems to me, is not whether such a computer-man can be constructed; it is rather why the idea is becoming so popular in a historical period when nothing seems to be more important than to transform the existing man into a more rational, harmonious, and peace-loving being. One cannot help being suspicious that often the attraction of the computer-man idea is the expression of a flight from life and from humane experience into the mechanical and purely cerebral (p.45).

 

The idea of the manlike computer is a good example of the alternative between the human and the inhuman use of machines. The computer can serve the enhancement of life in many respects. But the idea that it replaces man and life is the manifestation of the pathology of today (p.45-46).

Edited by william7, 24 October 2016 - 02:04 AM.

[mag1]

Another computer milestone reached.

Computers are now better at human speech recognition.

 

https://arxiv.org/pd...10.05256v1.pdf 

[mag1]

I am disappointed that the recent American election did not focus more on the imminent introduction of robotic transport.

 

Ignoring this issue will not somehow make it go away. From what I can see online, the awareness of this game changing

technology is growing, though there has as of yet been no organized social response. Waiting for the day when suddenly

200 million Americans become unemployed is not the best strategy to manage this impending crisis. Large tracts of our

urban/suburban landscape could soon be transformed into vast economic voids. While a reflexive "Smash the machines"

is entirely unpalatable, the complete silence on this issue is preventing even an initial effort of constructively thinking about

our future.  

[niner]

They couldn't even deal with climate change, where the solutions are relatively straightforward, so expecting them to deal with the coming robo/AI employment apocalypse is asking a lot.  It's a lot more comforting to blame job losses on "bad trade deals", but the majority of job losses in the past 15 years, and some 90% in the last decade have been a result of automation.  That trend will continue.  I wonder how long it will take downtrodden American middle class workers to realize that they've been sold a bill of goods by Mr. Trump?