They're definitely a pain. As a horse owner, I see too much of this, because the typical "horse age" is 12-16. (A joke in the horse business is that around 16 they discover boys and lose interest in horses, then around 35 they realize men are jerks and come back to horses.) So I get to see too much early teen angst.
Horse kids seem to get over it with less trouble. They're used to remaining calm around big, excitable animals. This keeps the drama level down.
One day, one of the high-school-aged teenagers at the Stanford barn came in after school and announced in frustration "I just don't understand teenage boys". I said "Go watch the rooster for fifteen minutes and you'll get it". The barn's rooster made a lot of noise, ran around a lot, tried to hump the hens, and accomplished very little. A rooster is testosterone with a peanut-sized brain. She watched, and said "Now I get it."
How many times must people be hit in the head with a clue bat before they understand that this is a Bad Idea[tm]
Big companies are encouraging this, by sending emails that meet all the criteria for phishing emails. I just got a receipt email from Virgin Mobile after making a payment. The path taken by the mail goes through "mh.nextel.m0.net", "oms16.dc1.prod" (which isn't even a valid TLD), and "cmil278.amdocs.com". The mail text is base-64 encoded HTML only, no text version.
That just screams "hostile code".
How are people supposed to recognize phishing emails with legit companies sending crap like that?
"m0.net" says on their site "This domain is owned by Acxiom Digital, a leading provider of email marketing solutions for Global 2000 enterprises."
From now on, all point-to-point message security has to be end to end. At no point in the middle can a message be plain text. The era of trusting service providers is over.
We really need is a good way for people to publish their public key, in a place where tampering with it will be detected. Somebody needs to solve that problem.
For example, Martin Luther King's speeches criticized the status quo. Since the status quo is now matter of National Security, Martin Luther King's speeches were a threat to National Security, by today's standards.
Nothing like the union movement or the black rights movement or the gay rights movement could happen today. It would be crushed. Back in the 1960s, there wasn't so much jail capacity, and cops were not well organized. So mass civil disobedience was possible. Now, if 10,000 people have to be sent to jail, no problem. Look what happened to the Occupy Wall Street movement.
Silicon Valley had that from 1974 to 1976. It was called "Dial a RIde". It was a popular service, but too expensive to provide. The hope was that there would be enough people going in roughly the same directions that the small buses used would fill up. But it turned out that there wasn't enough commonality of destination. Everybody wanted to go some place different, and the buses often had one passenger.
Most successful van systems have a common source or destination - a school or airport. Without some concentrating factor, cabs or cars are more effective.
Space-X is now able to provide much of the capability of the space shuttle at much lower prices. (Each Space Shuttle launch ended up costing about $600 million.) Once the Falcon Heavy launches, they'll have the capability for even more lift to LEO.
Ah. That's helpful. I'm seeing lots of stories about "loosening of control" from last summer, but capital flows are still highly restricted. As long as that's the case, the renminbi can't become an international reserve currency.
If exchange controls are lifted, the exchange rate between yuan and other currencies will no longer be under the control of the People's Bank of China. So far, maintaining control over exchange rates has been more important to China's leadership than being the issuer of the world's reserve currency.
This has possibilities for the last 100 meters of delivery - from the (soon self-driving) delivery truck to the customer. The truck stops near the destination, and a quadrotor takes the package to the door. The quadrotor only has to have a few minutes of battery life, since it gets recharged each time it returns to the truck. So it can trade power for endurance and carry more.
Apartment dwellers could have an air-conditioner sized landing pad outside their window for direct delivery.
China still has substantial currency controls on the renminbi. It's difficult to move renminbi to other currencies. There's a "State Administration of Foreign Exchange" which issues permits to do that. Businesses and individuals in China can buy goods with renminbi from other countries, but exchanging renminbi for dollars or euros or doing other cross-currency financial transactions is heavily controlled.
(Lately, there's been a surge in Bitcoin transactions in China. This provides a way to get around exchange controls. This activity will probably provoke some government action if it gets big.)
Because of those exchange controls, the renminbi has not been a major international currency. That was the deliberate policy of the People's Bank of China for years, because they didn't want their internal economy yanked around by external events. That policy changed in July 2013. China now has a big enough economy internally to outweigh external holders.
Retail controls are loosening. HSBC and Standard Chartered Bank will now let you open a bank account outside China denominated in yuan. But it's not freely exchangeable yet.
Even 100MHz RF will get through tens of meters of water. Ground penetrating radar systems can image through shallow water and into the ground underneath. The losses are huge, but remember that you can send very high power pulses and detect microwatts.
I have a mail filter that dumps anything with "Webinar" in the title to the spam folder. Most of them are infomercials. They're invariably too long for the information contained.
This, in fact, is the big problem with "online education" - no post production cutting.
"Regen paddles?" Why should the driver have to control the power train at that level? Regenerative braking should happen during any braking, as it does on most other electric cars. For light braking, regenerative braking is enough; push the brake pedal down further and the brake pads engage.
This isn't a new concept; it appeared first in the PCC streetcar, from 1936. (San Francisco still runs a fleet of them.) The PCC cars had a whole hierarchy of braking. As the brake pedal was depressed, first the drive motors went to regen mode, dumping power back into the trolley line or into big iron resistance grids on the roof. Then the brake shoes were applied to the wheels by compressed air. Next, four big rubber brake buffers pressed down against the rails at four points. Finally, if you floored the brake pedal, the sander came on and dropped sand ahead of the lead wheels for extra grip. The operator didn't have to think about this - just step on the brakes.
Sillier things have appeared in high-end electric cars. There was one European prototype which not only had "shifting", but an engine noise generator feeding the car speakers. You actually had to shift as speed increased. All this did was feed an input to the control electronics, but it gave the illusion that the driver was doing something "high performance".
Building that in Cupertino is a good thing for the city. It's a blah suburb.
Now here's a prestige research center - IBM Alamaden Research Center. That place produced several Nobel Prizes. It's on an isolated mountaintop. You drive for a mile after entering the property before reaching the buildings. The view from the cafeteria is of mountains, with no other buildings in sight.
Do you have any idea what's going on inside a modern vehicle stability control system? There are accelerometers and rate gyros, steering wheel sensors, wheel pulse counters on all four wheels, and driveshaft speed. All this feeds into a system which controls separately controllable brake actuators on all four wheels and can reduce the engine RPM. In a skid, the system tries to make the vehicle go in the direction the driver has requested via the steering wheel, without rolling the vehicle. This is non-trivial code.
It also has to self-test the hardware, detect and log faults, and not make driving worse due to a hardware fault.
No sane auto manufacturer is going to take the risk of legal liability. There will be accidents. And there will be lawsuits.
There's a financial engineering solution to that.
I am betting all the research that goes on at Ford or Toyota is just for the patents - they don't ever want to go in to production. Too risky.
That was true for a long time. GM played around with automatic driving as early as the 1950s.
It's not true any more. GM now intends to have full-auto driving in Cadillacs by 2016.
I wouldn't trust a web-oriented company with self-driving car development. They're into "agile","features now, quality later if ever" and "release and patch" development. That's not how avionics are developed, and production self-driving cars need avionics-quality software.
Self-driving needs the engineering discipline that comes from having to pay for a recall when it doesn't work, and paying for damages when it hurts someone.
Ah, Los Alamos. Once it had more great scientists in one place than anywhere else in the world. There was a tradition in the early days that the head of Los Alamos must have a Nobel Prize. That ended in the 1980s when Ronald Reagan put a lawyer in charge.
The US has a strange approach to "national laboratories". The original ones (Los Alamos, Lawerence Livermore, Sandia, Oak Ridge, Savannah River, etc.) were originally all Atomic Energy Commission operations. The Department of Energy got the AEC operations when it was formed. So the US still has a huge nuclear weapons R&D operation, despite the fact that the US hasn't built a new nuclear weapon in decades.
This project sounds more like an excuse for funding basic research than a component needed in a nuclear weapon. EMP shielding isn't that hard. This MEMS device doesn't seem to be a likely choice for the firing switch in a nuclear weapon. Nuclear weapons require a symmetrical implosion squeeze, which is initiated with multiple detonators, all of which have to go off at the same time within 1ns or so. This is done with a setup like a photoflash, but more powerful - a capacitor bank is charged up, and then dumped into thin wire detonators when the discharge switch closes. It's a few KV at a few thousand amps for a nanosecond or so. That discharge switch is what the article probably refers to.
The classic device for that is a krytron.
Although using a gas-filled tube is kind of retro, it works. It's probably possible to build some MOSFET device to replace krytrons, as this work at SLAC indicates. But a microscopic MEMS device? Too tiny to handle the current.
"1) Sharing content with people online is a poor substitute for having real-life experiences with, like, kayaking and family gatherings and drinking and stuff."
Is this sponsored content from the booze industry?
The booze industry puts massive efforts into making booze "cool". "Alcohol marketing can shape culture by creating and sustaining expectations and norms about how to achieve social, sporting or sexual success". It's not just commercials. There are discounts for areas near prestigious schools.
Haven't met him, so I can't say. But he does tend to issue press releases which are out of touch with reality. Most of them are of the form "Canonical teams with BIG_COMPANY to put Linux on MAJOR_PRODUCT". Somehow, when those Canonical press releases go out, there's never a corresponding press release from the BIG_COMPANY, and the product never materializes or shows up only as a low-volume niche thing.
So a Shuttleworth press release has little credibility.
Maybe not so much excitement. Our vehicles have already learned to maintain speed, enhance braking, honk-and-flash when the door is opened from the inside after the key's been out of the ignition too long, detect road obstacles and now can take over parking.
It's all incremental improvement, like the cinder-block-to-pocket-slate cell phone evolution.
There are already vehicle "autopilot" systems good enough to allow the driver to stop looking at the road some of the time. Radar-controlled cruise control plus lane keeping is enough to allow that. But it's not enough to prevent accidents caused by even slightly difficult situations. Several car companies have shipped "driving assist" systems which can do that, but they've deliberately kept them from operating with no driver input. Ford, Mercedes, BMW, and Audi have all stopped just before hands-off driving.
The auto industry recognizes that there is a "deadly valley" that begins when a vehicle "autopilot" is good enough to allow the driver to stop looking at the road some of the time. On the other side of the deadly valley is fully hands-off autonomous driving, which Google almost has now. We will see commercilaly successful systems on the other side of the deadly valley within the next decade.
Systems that operate in the "deadly valley" will make things worse, for obvious reasons.
What the W3C should demand in exchange for doing this is that all it does is prevent copying. Content can't "phone home". Content can't keep you from skipping the commercials. All this mechanism should do is restrict a container of content to one or more specified devices. It should not be used as a technical means to give content powers beyond those established by copyright law.
I think the pertinent question is whether Microsoft or Google or Yahoo should responsible for the ads they show.
That's a very good question. Because the major search engines do not vet their advertisers very well. Google had to pay $500,000,000 to the USDOJ when they were caught willfully running ads for an obvious drug dealer. (No, it wasn't about "Canadian pharmacies". Some Google apologists tried to spin it that way, but the details came out.)
Google has since clamped down. They had to; they were on DOJ probation for two years, with felony charges hanging over them. "Oxycontin no prescription" no longer returns ad results. Same for "viagra". Bing now pops up an "Is it legit?" box for searches like that.
Google's clampdown was narrow. Searches with "foreclosure" and "credit repair" have a high population of scammers. Financial search keywords carry a high price, because the marks can be taken for big amounts.
It's possible to measure basic advertiser legitimacy. We do that with SiteTruth, which tries to find the real-world business behind the ad. For over 30% of Google advertisers (by domain name), there's no identifiable real-world business behind the ad. (Running an anonymous business is illegal in some states and in the EU.) That's embarrassing, and highly profitable for Google.
Slashdot Mirror
They're definitely a pain. As a horse owner, I see too much of this, because the typical "horse age" is 12-16. (A joke in the horse business is that around 16 they discover boys and lose interest in horses, then around 35 they realize men are jerks and come back to horses.) So I get to see too much early teen angst.
Horse kids seem to get over it with less trouble. They're used to remaining calm around big, excitable animals. This keeps the drama level down.
One day, one of the high-school-aged teenagers at the Stanford barn came in after school and announced in frustration "I just don't understand teenage boys". I said "Go watch the rooster for fifteen minutes and you'll get it". The barn's rooster made a lot of noise, ran around a lot, tried to hump the hens, and accomplished very little. A rooster is testosterone with a peanut-sized brain. She watched, and said "Now I get it."
How many times must people be hit in the head with a clue bat before they understand that this is a Bad Idea[tm]
Big companies are encouraging this, by sending emails that meet all the criteria for phishing emails. I just got a receipt email from Virgin Mobile after making a payment. The path taken by the mail goes through "mh.nextel.m0.net", "oms16.dc1.prod" (which isn't even a valid TLD), and "cmil278.amdocs.com". The mail text is base-64 encoded HTML only, no text version. That just screams "hostile code".
How are people supposed to recognize phishing emails with legit companies sending crap like that?
"m0.net" says on their site "This domain is owned by Acxiom Digital, a leading provider of email marketing solutions for Global 2000 enterprises."
From now on, all point-to-point message security has to be end to end. At no point in the middle can a message be plain text. The era of trusting service providers is over.
We really need is a good way for people to publish their public key, in a place where tampering with it will be detected. Somebody needs to solve that problem.
For example, Martin Luther King's speeches criticized the status quo. Since the status quo is now matter of National Security, Martin Luther King's speeches were a threat to National Security, by today's standards.
Nothing like the union movement or the black rights movement or the gay rights movement could happen today. It would be crushed. Back in the 1960s, there wasn't so much jail capacity, and cops were not well organized. So mass civil disobedience was possible. Now, if 10,000 people have to be sent to jail, no problem. Look what happened to the Occupy Wall Street movement.
Silicon Valley had that from 1974 to 1976. It was called "Dial a RIde". It was a popular service, but too expensive to provide. The hope was that there would be enough people going in roughly the same directions that the small buses used would fill up. But it turned out that there wasn't enough commonality of destination. Everybody wanted to go some place different, and the buses often had one passenger.
Most successful van systems have a common source or destination - a school or airport. Without some concentrating factor, cabs or cars are more effective.
Space-X is now able to provide much of the capability of the space shuttle at much lower prices. (Each Space Shuttle launch ended up costing about $600 million.) Once the Falcon Heavy launches, they'll have the capability for even more lift to LEO.
Space-X even offers transportation of humans to low earth orbit. So far, NASA is the only buyer, but Space-X advertises it as a commercial service.
Ah. That's helpful. I'm seeing lots of stories about "loosening of control" from last summer, but capital flows are still highly restricted. As long as that's the case, the renminbi can't become an international reserve currency.
If exchange controls are lifted, the exchange rate between yuan and other currencies will no longer be under the control of the People's Bank of China. So far, maintaining control over exchange rates has been more important to China's leadership than being the issuer of the world's reserve currency.
This has possibilities for the last 100 meters of delivery - from the (soon self-driving) delivery truck to the customer. The truck stops near the destination, and a quadrotor takes the package to the door. The quadrotor only has to have a few minutes of battery life, since it gets recharged each time it returns to the truck. So it can trade power for endurance and carry more.
Apartment dwellers could have an air-conditioner sized landing pad outside their window for direct delivery.
China still has substantial currency controls on the renminbi. It's difficult to move renminbi to other currencies. There's a "State Administration of Foreign Exchange" which issues permits to do that. Businesses and individuals in China can buy goods with renminbi from other countries, but exchanging renminbi for dollars or euros or doing other cross-currency financial transactions is heavily controlled.
(Lately, there's been a surge in Bitcoin transactions in China. This provides a way to get around exchange controls. This activity will probably provoke some government action if it gets big.)
Because of those exchange controls, the renminbi has not been a major international currency. That was the deliberate policy of the People's Bank of China for years, because they didn't want their internal economy yanked around by external events. That policy changed in July 2013. China now has a big enough economy internally to outweigh external holders.
Retail controls are loosening. HSBC and Standard Chartered Bank will now let you open a bank account outside China denominated in yuan. But it's not freely exchangeable yet.
Here's a summary of "Circular Concerning the Simplification of Cross-Border RMB Procedures and Improvement of Relevant Policies" from the People's Bank of China. The changes are slow and cautious, but are happening.
Even 100MHz RF will get through tens of meters of water. Ground penetrating radar systems can image through shallow water and into the ground underneath. The losses are huge, but remember that you can send very high power pulses and detect microwatts.
I have a mail filter that dumps anything with "Webinar" in the title to the spam folder. Most of them are infomercials. They're invariably too long for the information contained.
This, in fact, is the big problem with "online education" - no post production cutting.
Central Planning does NOT work.
Oh? Check out Walmart, McDonalds, Apple...
"Regen paddles?" Why should the driver have to control the power train at that level? Regenerative braking should happen during any braking, as it does on most other electric cars. For light braking, regenerative braking is enough; push the brake pedal down further and the brake pads engage.
This isn't a new concept; it appeared first in the PCC streetcar, from 1936. (San Francisco still runs a fleet of them.) The PCC cars had a whole hierarchy of braking. As the brake pedal was depressed, first the drive motors went to regen mode, dumping power back into the trolley line or into big iron resistance grids on the roof. Then the brake shoes were applied to the wheels by compressed air. Next, four big rubber brake buffers pressed down against the rails at four points. Finally, if you floored the brake pedal, the sander came on and dropped sand ahead of the lead wheels for extra grip. The operator didn't have to think about this - just step on the brakes.
Sillier things have appeared in high-end electric cars. There was one European prototype which not only had "shifting", but an engine noise generator feeding the car speakers. You actually had to shift as speed increased. All this did was feed an input to the control electronics, but it gave the illusion that the driver was doing something "high performance".
Building that in Cupertino is a good thing for the city. It's a blah suburb.
Now here's a prestige research center - IBM Alamaden Research Center. That place produced several Nobel Prizes. It's on an isolated mountaintop. You drive for a mile after entering the property before reaching the buildings. The view from the cafeteria is of mountains, with no other buildings in sight.
It's also half-empty since IBM cut back.
Automotive software is mostly pretty boring.
Do you have any idea what's going on inside a modern vehicle stability control system? There are accelerometers and rate gyros, steering wheel sensors, wheel pulse counters on all four wheels, and driveshaft speed. All this feeds into a system which controls separately controllable brake actuators on all four wheels and can reduce the engine RPM. In a skid, the system tries to make the vehicle go in the direction the driver has requested via the steering wheel, without rolling the vehicle. This is non-trivial code. It also has to self-test the hardware, detect and log faults, and not make driving worse due to a hardware fault.
Such code is not written by PHP programmers.
No sane auto manufacturer is going to take the risk of legal liability. There will be accidents. And there will be lawsuits.
There's a financial engineering solution to that.
I am betting all the research that goes on at Ford or Toyota is just for the patents - they don't ever want to go in to production. Too risky.
That was true for a long time. GM played around with automatic driving as early as the 1950s. It's not true any more. GM now intends to have full-auto driving in Cadillacs by 2016.
I wouldn't trust a web-oriented company with self-driving car development. They're into "agile","features now, quality later if ever" and "release and patch" development. That's not how avionics are developed, and production self-driving cars need avionics-quality software.
Self-driving needs the engineering discipline that comes from having to pay for a recall when it doesn't work, and paying for damages when it hurts someone.
Ah, Los Alamos. Once it had more great scientists in one place than anywhere else in the world. There was a tradition in the early days that the head of Los Alamos must have a Nobel Prize. That ended in the 1980s when Ronald Reagan put a lawyer in charge.
The US has a strange approach to "national laboratories". The original ones (Los Alamos, Lawerence Livermore, Sandia, Oak Ridge, Savannah River, etc.) were originally all Atomic Energy Commission operations. The Department of Energy got the AEC operations when it was formed. So the US still has a huge nuclear weapons R&D operation, despite the fact that the US hasn't built a new nuclear weapon in decades.
This project sounds more like an excuse for funding basic research than a component needed in a nuclear weapon. EMP shielding isn't that hard. This MEMS device doesn't seem to be a likely choice for the firing switch in a nuclear weapon. Nuclear weapons require a symmetrical implosion squeeze, which is initiated with multiple detonators, all of which have to go off at the same time within 1ns or so. This is done with a setup like a photoflash, but more powerful - a capacitor bank is charged up, and then dumped into thin wire detonators when the discharge switch closes. It's a few KV at a few thousand amps for a nanosecond or so. That discharge switch is what the article probably refers to.
The classic device for that is a krytron. Although using a gas-filled tube is kind of retro, it works. It's probably possible to build some MOSFET device to replace krytrons, as this work at SLAC indicates. But a microscopic MEMS device? Too tiny to handle the current.
"1) Sharing content with people online is a poor substitute for having real-life experiences with, like, kayaking and family gatherings and drinking and stuff."
Is this sponsored content from the booze industry?
The booze industry puts massive efforts into making booze "cool". "Alcohol marketing can shape culture by creating and sustaining expectations and norms about how to achieve social, sporting or sexual success". It's not just commercials. There are discounts for areas near prestigious schools.
Is Shuttleworth fucking stupid?
Haven't met him, so I can't say. But he does tend to issue press releases which are out of touch with reality. Most of them are of the form "Canonical teams with BIG_COMPANY to put Linux on MAJOR_PRODUCT". Somehow, when those Canonical press releases go out, there's never a corresponding press release from the BIG_COMPANY, and the product never materializes or shows up only as a low-volume niche thing.
So a Shuttleworth press release has little credibility.
Maybe not so much excitement. Our vehicles have already learned to maintain speed, enhance braking, honk-and-flash when the door is opened from the inside after the key's been out of the ignition too long, detect road obstacles and now can take over parking.
It's all incremental improvement, like the cinder-block-to-pocket-slate cell phone evolution.
There are already vehicle "autopilot" systems good enough to allow the driver to stop looking at the road some of the time. Radar-controlled cruise control plus lane keeping is enough to allow that. But it's not enough to prevent accidents caused by even slightly difficult situations. Several car companies have shipped "driving assist" systems which can do that, but they've deliberately kept them from operating with no driver input. Ford, Mercedes, BMW, and Audi have all stopped just before hands-off driving.
The auto industry recognizes that there is a "deadly valley" that begins when a vehicle "autopilot" is good enough to allow the driver to stop looking at the road some of the time. On the other side of the deadly valley is fully hands-off autonomous driving, which Google almost has now. We will see commercilaly successful systems on the other side of the deadly valley within the next decade.
Systems that operate in the "deadly valley" will make things worse, for obvious reasons.
What the W3C should demand in exchange for doing this is that all it does is prevent copying. Content can't "phone home". Content can't keep you from skipping the commercials. All this mechanism should do is restrict a container of content to one or more specified devices. It should not be used as a technical means to give content powers beyond those established by copyright law.
I think the pertinent question is whether Microsoft or Google or Yahoo should responsible for the ads they show.
That's a very good question. Because the major search engines do not vet their advertisers very well. Google had to pay $500,000,000 to the USDOJ when they were caught willfully running ads for an obvious drug dealer. (No, it wasn't about "Canadian pharmacies". Some Google apologists tried to spin it that way, but the details came out.) Google has since clamped down. They had to; they were on DOJ probation for two years, with felony charges hanging over them. "Oxycontin no prescription" no longer returns ad results. Same for "viagra". Bing now pops up an "Is it legit?" box for searches like that.
Google's clampdown was narrow. Searches with "foreclosure" and "credit repair" have a high population of scammers. Financial search keywords carry a high price, because the marks can be taken for big amounts.
It's possible to measure basic advertiser legitimacy. We do that with SiteTruth, which tries to find the real-world business behind the ad. For over 30% of Google advertisers (by domain name), there's no identifiable real-world business behind the ad. (Running an anonymous business is illegal in some states and in the EU.) That's embarrassing, and highly profitable for Google.
Misread the article. It's just Glenn now, the last of the Original Seven human astronauts.
Glenn is still alive. Who's the other one?