They have to solve the processing problem. It's really hard to find automotive grade processors that can run safety critical code with enough performance for these applications. They need to be programmable and flexible. They need to handle the harsh automotive environment. There needs to be significant redundancies. It also need to be cheap.
The complexity of the systems are orders of magnitude more complicated than anything automotive companies have dealt with in the past. Automotive companies are not the best when it comes to complicated software systems. They tend to treat everything as an isolated piece of hardware and spin off custom protocols for everything.
Google uses whatever hardware they want because they expect a person to be there. When your start making hundreds of cars, you have to build them to automotive standards.
Face recognition these days uses a lot of deep learning. Deep learning does not use "biometric identifiers" or "face geometry", it operates on the images directly.
It's funny that Deep learning has turned into this new buzzword. It helps to use it correctly. I doubt that most face recognition uses deep learning since deep convolutional neural networks require a large training set to train them. Additionally, they would have issues handling views that have never been seen before. Deep learning is a type of machine learning, but there are other types as well. Read the OpenCV tutorial on face detection. None of those examples use deep learning.
Even when using machine learning you are using "biometric identifiers" and "face geometry", albeit indirectly. Sometimes you use it directly. Some methods look for eyes, nose, corners of the mouth and use those as inputs for recognition.
Depending on where you live in the country, churches can have a much larger role in the community. I think something like 83% of Americans identify as Christians.
It's worse than that. Some drones will land themselves, but drones currently don't have a requirement for backup plans if they lose links, so if you jam a harmless drone it may go out of control and kill someone.
Additionally, there are other legitimate uses for the bands that drones use. The FCC wouldn't be happy about airplanes disrupting those services.
Drones also can be controller by pretty much any radio signal including a cell phone signal. The FCC REALLY wouldn't be happy about jamming cell phones.
Most importantly, most drones have some ability to fly autonomously even though they are supposed to be piloted by an actual person on the ground. It is concievable that any malicious drone would hit it's target without any pilot. It's not that hard to program a drone with custom sensors and firmware.
I'd use a typewriter and a carbon paper for backup.
The problem with paper is that it has the same problem of storing a floppy in a box for 30 years. Floppies and paper degrade over time. Paper degrades slower, but if you put the box in a bad location, it's not going to last very long.
the CP/M disks were far more work
cpmtools looks fairly easy to use and easy to install on ubuntu. sudo apt-get install cpmtools You can probably get it all done with a dd and a cpmcp command.
I think the corrupt disks would be the trickiest part.
That is correct. Then just let lawyers decide which ones are enforceable once they are challenged. I laugh every time I hear someone brag about their patent portfolio before it's been properly vetted by those blood suckers.
If it's a choice between all or nothing, then I'd pick nothing.
Port scan alerts are a bad idea for three reasons. 1. These attacks are very common and excess noise of the alerts may distract you from real threats. 2, Port scans that get caught by these filters are usually benign. NMAP is the first tool that every little kid who thinks they are a hacker plays with before they learn some common sense. 3. Any sophisticated attack that actually stands a chance of working won't be detected by these simple mechanisms.
Hopefully, your firewall will detect the real threats using more sophisticated methods. If I were you I wouldn't count on it catching everything. Those alerts might be giving you a false sense of security. The only thing that alert is satisfying is the author's curiosity. It's not really protecting him.
I agree. How hard could it possibly be to build a glorified seismograph?
I know everything must be space rated and communications are involved, but these should be problems with well defined solutions. I feel as though they try to get clever when it comes to space stuff and it never works out quite the way they expect.
The sad thing is the company gets paid whether they launch or not.
Silver is an element. If you recycle your electronics it can be recovered. when it becomes scarce enough to care we'll just melt down our trash and get the precious metals out of it.
It's difficult to make an ink that comes close to the conductivity of copper used in pcbs. It's even more difficult to make it stretch. Silver is a great conductor. We use gold because it doesn't corrode, but it sounds like silver has some desired properties that would be difficult to reproduce with another metal.
Once Torrent is served it is immutable. You cannot add things to it. The torrent could be infected during creation, but what really works is infecting the site that you use to download the torrent file. The easiest method is to trick users into thinking they need a browser plugin or something.
More importantly the originally raspi is based on a chip that was intended for a different purpose. The bcm2835 was first and foremost a video processor capable of hd video encode and decode.
The arm(which everyone benchmarks) has a simple role to play in the intended configuration: Run linux, so that you can write simple GUIs and send compressed video data to the videocore. 3d Graphics acceleration was probably and afterthought.
You're comparing apples and oranges because the raspi community repurposed the chip. The raspi2 arm is a bit better. The biggest advantage now is the price, but this comparison is ridiculous.
Exactly, overt prevention discourages use. You get way better intelligence by allowing use of these resources. Additionally any attempt to disrupt use will inevitably prevent some legitimate use.
The article suggests a new mode of driving only possible by driverless cars, but 1 meter apart is kinda ridiculous. A tire failure in any one of those cars could cause a pileup of unimaginable purportions. I'd settle for autonomous cars driving at human following distances because we know humans can do it. Even at human following distances autonomous cars can improve things because even the simplest actions on the road have huge unseen consequences from the drivers perspective. You can avoid the problem where a dumb driver can tap on his brakes for no good reason and sends ripples of break lights miles behind him/her. Sometimes I wonder how many of those drivers drive with two feet. One on the brake and one on the accelerator.
The problem with teaching someone to program is that they tend to miss the part where you have to think about the problem abstractly and structure the problem properly. If you can communicate the problem effectively to other people, then you can find a programmer to code up your idea relatively easily. The aptitude problem comes from the fact that many people think learning to program is the point. Good programmers are able to communicate abstract concepts at a high level, which can't really be taught through just learning to program. Honestly, I think you learn how to program better simply by explaining what you are trying to do in plain English. The actual programming task has almost little educational value.
Students should come up with a problem they are trying to solve, conceptualize it and then explain it to their coding buddy and if their coding buddy can code it then the student probably understands how to program well.
I can't imagine this being useful for a post office in developed countries
Well, think this is an example from TFA (Japanese characters removed):
Here is just one example, an address in Tokyo. 100-8994 (zip code), (Tokyo-to, i.e. Tokyo prefecture or state) (Chuo-ku, i.e. Chuo Ward) (Yaesu 1-chome, i.e. Yaesu district 1st subdistrict) (block 5 lot 3), (Tokyo Central Post Office).
Apparently, in some places addresses can get pretty screwed up.
Address systems are difficult to change. If it was easy to change a system then all the Japanese addresses would be changed to something sane. The hierarchical address designation in the in the US follows the same uniform format almost everywhere and seems to be simple enough for most people to follow. I think most people prefer this to what is more or less a GPS location. The hierarchy is typically, country->state->city->road->address. Although we get screwy with zip codes. Some cities have multiple zip codes and some cities share zip codes, so feel free to replace city with zip. It's funny though that zip codes are more designed for processing the mail, which much of it is done by machines now. I still think this structure is far better for post offices because post offices interact with people. Machines can care less about these things that humans use to organize locations. Finding a postion based on GPS location is much simpler for a machine. Even if machines use roads the mapping GPS positions to roads is quite trivial. The 3 word combination is more for the machines than it is the humans.
You're missing the point. Words are easier to remember. IP addresses are often shorter than most domain names, but we still use domain names because they can be easily remembered.
news.slashdot.org uses more characters than the ip address 216.34.181.48. You can even remove the dots in the ip address and save some space. Or better yet you can encode the ip address as hex. Now you have an 8 character string vs an 18 character string. I still prefer news.slashdot.org.
I can't imagine this being useful for a post office in developed countries. Drones on the other hand, are going to deliver packages in a back yard and if you can tell the drone search for a place to drop a package in a 3m by 3m square that's definately useful. Especially if there is a designator nearby to better pinpoint the landing zone.
This is how crony capitalism works. Perverse incentives in the tax code are exploited, so that rich insiders can avoid paying fair shares in taxes. Not saying I agree with significantly increasing taxes on the rich, but it does make it harder to break past that glass ceiling if you don't know how to strategize around unnecessarily complex tax laws which actually impose unnecessarily economic costs of their own in the grand scheme of things.
I mean some people's tax rates are just criminally low even after deducting charitable contributions from their income.
Slashdot Mirror
They have to solve the processing problem. It's really hard to find automotive grade processors that can run safety critical code with enough performance for these applications. They need to be programmable and flexible. They need to handle the harsh automotive environment. There needs to be significant redundancies. It also need to be cheap.
The complexity of the systems are orders of magnitude more complicated than anything automotive companies have dealt with in the past. Automotive companies are not the best when it comes to complicated software systems. They tend to treat everything as an isolated piece of hardware and spin off custom protocols for everything.
Google uses whatever hardware they want because they expect a person to be there. When your start making hundreds of cars, you have to build them to automotive standards.
I only made my system quindecuply redundant. Why oh why didn't I make it sexdecuply redundant.
Face recognition these days uses a lot of deep learning. Deep learning does not use "biometric identifiers" or "face geometry", it operates on the images directly.
It's funny that Deep learning has turned into this new buzzword. It helps to use it correctly. I doubt that most face recognition uses deep learning since deep convolutional neural networks require a large training set to train them. Additionally, they would have issues handling views that have never been seen before. Deep learning is a type of machine learning, but there are other types as well. Read the OpenCV tutorial on face detection. None of those examples use deep learning.
Even when using machine learning you are using "biometric identifiers" and "face geometry", albeit indirectly. Sometimes you use it directly. Some methods look for eyes, nose, corners of the mouth and use those as inputs for recognition.
Depending on where you live in the country, churches can have a much larger role in the community. I think something like 83% of Americans identify as Christians.
It's worse than that.
Some drones will land themselves, but drones currently don't have a requirement for backup plans if they lose links, so if you jam a harmless drone it may go out of control and kill someone.
Additionally, there are other legitimate uses for the bands that drones use. The FCC wouldn't be happy about airplanes disrupting those services.
Drones also can be controller by pretty much any radio signal including a cell phone signal. The FCC REALLY wouldn't be happy about jamming cell phones.
Most importantly, most drones have some ability to fly autonomously even though they are supposed to be piloted by an actual person on the ground. It is concievable that any malicious drone would hit it's target without any pilot. It's not that hard to program a drone with custom sensors and firmware.
Because there aren't enough North Koreans to throw the damn missile.
Why worry?
I'd use a typewriter and a carbon paper for backup.
The problem with paper is that it has the same problem of storing a floppy in a box for 30 years. Floppies and paper degrade over time. Paper degrades slower, but if you put the box in a bad location, it's not going to last very long.
the CP/M disks were far more work
cpmtools looks fairly easy to use and easy to install on ubuntu.
sudo apt-get install cpmtools
You can probably get it all done with a dd and a cpmcp command.
I think the corrupt disks would be the trickiest part.
That is correct. Then just let lawyers decide which ones are enforceable once they are challenged. I laugh every time I hear someone brag about their patent portfolio before it's been properly vetted by those blood suckers.
If it's a choice between all or nothing, then I'd pick nothing.
Port scan alerts are a bad idea for three reasons.
1. These attacks are very common and excess noise of the alerts may distract you from real threats.
2, Port scans that get caught by these filters are usually benign. NMAP is the first tool that every little kid who thinks they are a hacker plays with before they learn some common sense.
3. Any sophisticated attack that actually stands a chance of working won't be detected by these simple mechanisms.
Hopefully, your firewall will detect the real threats using more sophisticated methods. If I were you I wouldn't count on it catching everything. Those alerts might be giving you a false sense of security. The only thing that alert is satisfying is the author's curiosity. It's not really protecting him.
I agree. How hard could it possibly be to build a glorified seismograph?
I know everything must be space rated and communications are involved, but these should be problems with well defined solutions. I feel as though they try to get clever when it comes to space stuff and it never works out quite the way they expect.
The sad thing is the company gets paid whether they launch or not.
I'd let you put it in my backyard if you let me re-refine and sell it back to you for your next gen reactors.
Nuclear fuel is recyclable if you have the right reactors.
Silver is an element. If you recycle your electronics it can be recovered. when it becomes scarce enough to care we'll just melt down our trash and get the precious metals out of it.
Doesn't take much silver to make an electrical connection.
It's difficult to make an ink that comes close to the conductivity of copper used in pcbs. It's even more difficult to make it stretch. Silver is a great conductor. We use gold because it doesn't corrode, but it sounds like silver has some desired properties that would be difficult to reproduce with another metal.
Don't forget to turn off cookies too.
Google knows.
Once Torrent is served it is immutable. You cannot add things to it. The torrent could be infected during creation, but what really works is infecting the site that you use to download the torrent file. The easiest method is to trick users into thinking they need a browser plugin or something.
More importantly the originally raspi is based on a chip that was intended for a different purpose. The bcm2835 was first and foremost a video processor capable of hd video encode and decode.
The arm(which everyone benchmarks) has a simple role to play in the intended configuration: Run linux, so that you can write simple GUIs and send compressed video data to the videocore. 3d Graphics acceleration was probably and afterthought.
You're comparing apples and oranges because the raspi community repurposed the chip. The raspi2 arm is a bit better. The biggest advantage now is the price, but this comparison is ridiculous.
Exactly, overt prevention discourages use. You get way better intelligence by allowing use of these resources. Additionally any attempt to disrupt use will inevitably prevent some legitimate use.
The article suggests a new mode of driving only possible by driverless cars, but 1 meter apart is kinda ridiculous. A tire failure in any one of those cars could cause a pileup of unimaginable purportions. I'd settle for autonomous cars driving at human following distances because we know humans can do it. Even at human following distances autonomous cars can improve things because even the simplest actions on the road have huge unseen consequences from the drivers perspective. You can avoid the problem where a dumb driver can tap on his brakes for no good reason and sends ripples of break lights miles behind him/her. Sometimes I wonder how many of those drivers drive with two feet. One on the brake and one on the accelerator.
The problem with teaching someone to program is that they tend to miss the part where you have to think about the problem abstractly and structure the problem properly. If you can communicate the problem effectively to other people, then you can find a programmer to code up your idea relatively easily. The aptitude problem comes from the fact that many people think learning to program is the point. Good programmers are able to communicate abstract concepts at a high level, which can't really be taught through just learning to program. Honestly, I think you learn how to program better simply by explaining what you are trying to do in plain English. The actual programming task has almost little educational value.
Students should come up with a problem they are trying to solve, conceptualize it and then explain it to their coding buddy and if their coding buddy can code it then the student probably understands how to program well.
Well, think this is an example from TFA (Japanese characters removed):
Apparently, in some places addresses can get pretty screwed up.
Address systems are difficult to change. If it was easy to change a system then all the Japanese addresses would be changed to something sane. The hierarchical address designation in the in the US follows the same uniform format almost everywhere and seems to be simple enough for most people to follow. I think most people prefer this to what is more or less a GPS location. The hierarchy is typically, country->state->city->road->address. Although we get screwy with zip codes. Some cities have multiple zip codes and some cities share zip codes, so feel free to replace city with zip. It's funny though that zip codes are more designed for processing the mail, which much of it is done by machines now. I still think this structure is far better for post offices because post offices interact with people. Machines can care less about these things that humans use to organize locations. Finding a postion based on GPS location is much simpler for a machine. Even if machines use roads the mapping GPS positions to roads is quite trivial. The 3 word combination is more for the machines than it is the humans.
You're missing the point. Words are easier to remember. IP addresses are often shorter than most domain names, but we still use domain names because they can be easily remembered.
news.slashdot.org uses more characters than the ip address 216.34.181.48. You can even remove the dots in the ip address and save some space. Or better yet you can encode the ip address as hex. Now you have an 8 character string vs an 18 character string. I still prefer news.slashdot.org.
I can't imagine this being useful for a post office in developed countries. Drones on the other hand, are going to deliver packages in a back yard and if you can tell the drone search for a place to drop a package in a 3m by 3m square that's definately useful. Especially if there is a designator nearby to better pinpoint the landing zone.
This is how crony capitalism works. Perverse incentives in the tax code are exploited, so that rich insiders can avoid paying fair shares in taxes. Not saying I agree with significantly increasing taxes on the rich, but it does make it harder to break past that glass ceiling if you don't know how to strategize around unnecessarily complex tax laws which actually impose unnecessarily economic costs of their own in the grand scheme of things.
I mean some people's tax rates are just criminally low even after deducting charitable contributions from their income.