Posted on 09/16/2019 8:23:01 PM PDT by lasereye
When will automotive industry investors learn the lesson from this weekend's reported drone attack on a Saudi oil facility?
We saw one version of this kind of scenario in the 1977 movie "Telefon" with Charles Bronson: Brainwashed sleeper cells were sent out in trucks to smash into various critical facilities, causing huge explosions.
The attack on the Saudi oil facility explains why it's illegal to operate a drone near an airport or a city center anywhere in America today. The risk of terrorist sabotage is simply too high.
But what is a driverless car, if not a 5,000-pound, land-based drone? We saw it not only in "Telefon" but also in car/truck bomb events in Afghanistan and Iraq over the last 10+ years.
Automakers can crow all they want about their intent to make their future driverless vehicles unhackable. Gee, why didn't anyone think of making a connected computer unhackable before? Please, someone call Microsoft, Apple, Google, Dell, Hewlett-Packard or any of the other computer industry entities and inform them about this revolutionary -- and easy-to-implement -- idea.
The fact is, there is no such thing as an unhackable connected computer, and there will never be. Thinking that one can be created is as futile as attempting to ban human stupidity. It's in the nature of things, and can't be wished away.
It's not only about driverless cars being hacked. Numerous Chinese brands have been salivating for years to export their soon-to-be driverless cars to the U.S. Those cars are "electric and connected" today -- and would become "driverless" when the software has "been perfected."
Then, they will be "turned on" -- or remotely upgraded -- to full driverless capability. Sort of like the sleeper cell telephone call in "Telefon."
You ought to be able to conclude from this that neither driverless cars -- of any origin -- nor Chinese cars of any kind, will ever be allowed to be sold or operated in the U.S. They constitute an even graver threat to U.S. national security than the presence of a Huawei switch in a U.S. telecom network such as AT&T or Verizon.
The message for the world's automakers and their investors is clear: When it comes to driverless cars, give it up. It's a futile exercise. If they are ever made to work, they must be immediately banned for national security reasons.
This weekend's drone attack in Saudi Arabia is yet another piece of evidence of why this is the only logical conclusion. A driverless car is just a very large land-based drone that can carry a lot more weight.
For all the automakers and other technology companies who are pursuing the pathway to driverless cars, ranging from Alphabet (GOOGL - Get Report) , Ford (F - Get Report) , Tesla (TSLA - Get Report) and General Motors (GM - Get Report) just to mention some of the more prominent ones, the futility of this outcome points to a multi-billion dollar write-off in the end.
The question is only how quickly such a write-off will have to be realized -- will it be closer to two years from now, or ten years from now?
There is a remote driver, its just that the instructions for control have been recorded in the form of computer code. The person, persons or entity that records the code is in control of the vehicle and thus driving. The only question is can you retake control of the vehicle when you so choose? Right now the answer is probably yes, in time definitely not.
The remote driver is the one who wrote the software, and the individual who wrote the requirements for the software.
You seem confused.
This isn’t a remote control system.
There isn’t anyone looking at a screen deciding which way the car should go.
This is literally a set of cameras sending video data to the on-board computer (actually a massive “artificial neural network”), which absolutely is making all the driving decisions on its own.
The car is, literally, self-driving - no human making decisions along the way.
re: “ control have been recorded in the form of computer code.”
Out of your depth ...
Well of course it’s a man-made machine. This did not evolve on its own, nor did God breathe life into it.
It does, however, drive itself.
And - here’s the wild part - it learned to.
Yes, it could have been programmed to obey the whims of a programmer or remote operator. It wasn’t. It was developed to go to where the driver indicates. The fleet of cars (in Tesla’s case) collaborated to use machine learning to teach the computers how to drive there: the computers basically observed people driving in real world scenarios, and learned to drive ... there is no explicit software coded to handle specific issues.
If you’re going to quibble over this still being “human is driving” you’re redefining terms to fit an agenda that can’t be discussed with and has no point.
In the case of “machine learning”, I would agree that there is not a human behind the behavior patterns, directly.
I would, however, be cautious about what the machine is learning. It often has only a tangential relationship to the goal being pursued.
Before Desert Storm, there was a neural network set up to identify camouflaged tanks on the battlefield. It became proficient, using satellite photos from Europe (from field exercises). When used on Desert Storm photos, it failed abysmally. It had learned to count the number of leaves, because in Europe the camouflage was made to look like foliage.
In the case of “machine learning”, I have serious concerns about what the machine is learning, and how the constraints on that learning affect the machine’s ability to perform. In aircraft, the pilot isn’t there for routine conditions - they are there because sometimes the computers can’t cope, or they break. I’m concerned that autonomous cars using self-learned behavior will be unable to react to stimuli sufficiently different from their learned inputs that the program cannot decide on an action.
(I know I changed the substance of my previous post - but did so in response to the information in yours. Forgive the faux pas, please.)
May I suggest you carve out some time and watch the Tesla Autonomy Day video: https://youtu.be/Ucp0TTmvqOE
Yes, it’s 4 hours long.
Yes, it may not be exciting.
Yes, the petty “but what about”s have been addressed.
It does make clear that Tesla is using full blown neural network machine learning on a scale way beyond anything ever attempted, and is wildly succeeding.
The cars are, in every sensible stretch of the term, driving themselves.
Because for the most part evil people aren't creative or intelligent. Most terrorists - there like here (antifa) are the loser children of powerful wealthy parents.
Now that the drone way of destroying is known - the lessor ones will copycat it. It's what losers do.
One breakthrough for Tesla: the entire fleet collaborates. When one encounters a new/unusual event, they all learn from it. It’s not just your car learning, it’s all of them - a hive mind, to use the applicable scary sci-fi term. Really unusual events are flagged and sent to humans to review, refine, and issue special updates (say, when a car flips over in front of one and the fleet can now learn what the bottom of a car looks like and associated behaviors).
Yes there will be some bizarre & disturbing special cases (like the recent video of a cyclist getting run over (uhh, don’t leap out in front of oncoming vehicles in dark areas)). Overall, injuries & deaths from self driving cars are substantially less than human drivers.
Thanks, but no.
Tesla’s cars may drive themselves, but not with me on board.
And hopefully not in proximity - especially on my motorcycle.
re: “Because for the most part evil people aren’t creative or intelligent. Most terrorists - there like here (antifa) are the loser children of powerful wealthy parents.
Now that the drone way of destroying is known - the lessor ones will copycat it. It’s what losers do. “
NOT exactly “stopping” them now is it?
You’re appealing to the fact that other factors enter into the final decision-making process, and maybe that also takes place in their minds.
So, why would the introduction of driverless cars change the dynamic? You just admitted “evil people aren’t creative or intelligent”.
I’m less sanguine on driverless cars, but I think we can agree to disagree on this. Thank you for engaging in an informed debate (even should you see my side as uninformed - LOL).
You’re at much greater risk on your bike with humans driving nearby.
In 5 years auto-drive will be common.
In 10 years auto-drive may be mandatory.
Such I’ve seen happen with many “thanks but no” technologies.
My grandfather lamented that my kids would grow up thinking computers were normal.
My daughter’s first word was “iPad”.
I expect her kids will learn driving only as an arcane skill.
That argument makes sense only if you assume the terrorists can develop their own driverless cars.
Why are you two arguing?
Self driving cars are happening.
Some jerk is going to use one for bomb delivery.
Not hard, yes we’re going to suffer from it.
That’s not going to stop proliferation of driverless cars, due to staggering convenience and inevitability of technology.
Drones are cheaper than driverless cars...
20 plus years in IT, so not really. What you and many others seem to have trouble accepting is that there will always be humans behind the control instructions of these vehicles and you are not one of them. Bottom line you are riding in a high tech taxi cab controlled by people other than you. Thats not “self driving”, it’s someone else driving and you are a passenger. I and many others see the existential danger to liberty this imposes.
re: “20 plus years in IT, so not really.”
Not the same thing as writing code, solving engineering problems.
We used to call the ppl who worked in the computer center “operators”, they changed/mounted tapes on the drives per user request, etc. ... somewhere along the time the title changed to “IT Professional” ... job responsibilities increased some, but don’t confuse yourself with those who actually design systems for a living ...
re: “Thats not “self driving”, it’s someone else driving”
Sadly, not true; It’s “something” doing the driving, an un-animated ‘something’ that operates within prescribed limits.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.