However, I will admit that this is certainly beyond the capabilities of most people due to the lack test equipment that is needed to even test the parts found at the swap meet.
I'm impressed with what this guy found at a hamfest. We don't see much microwave gear in Silicon Valley surplus any more. eBay, though, has a decent selection of microwave horns, low noise amplifiers, mixers, and waveguide. It looks like anybody could get the necessary parts in small quantity. New, though, those parts are expensive, so building low-cost robot vision systems this way is hard.
Also, when your "garage machine shop" has a Bridgeport milling machine, you're way above the usual home shop level. Still, if there's a TechShop in your town, you can get access to such machines.
A big problem working in this area, even if you know what you're doing, is that the test gear you need costs more than the thing you're making. Reading the design notes, some of which are on Air Force Research Lab stationery, indicate that the hamfest parts were tested and characterized using reasonably good test gear. And this was an MIT student, with access to MIT labs.
I ran into that building a small LIDAR in the early 1990s. The parts cost wasn't too bad, but I needed access to about $20K in test gear to debug the thing.
It's a real problem for Zuckerman. He's previously made fund-raising trips to Dubai. That's over. The UAE has blasphemy laws, which they enforce. The UAE also has an extradition treaty with Pakistan, but not with the United States. So he can no longer visit Dubai, and is unlikely to get funding from any source in the Arab world. He can't even fly Emirates Air.
Computer chess is merely at the point that if you haven't been on the cover of Chess Life, you're going to get trounced. Even if you have, you're going to lose more than you win. The current situation is that Deep Rybka 2010 has an ELO rating around 3150. That's running on a 4-core AMD-64 desktop machine. The all-time human record is 2851, which Garry Kasparov had in 1999-2000.
All this "smart meter" stuff has too much data flowing around, and too much of it is sent out of the home back to organizations which may make dubious use of it.
All that's really needed to level out peaks is to broadcast a few bits of information per hour to interested power-using devices. Here's the California Independent System Operator status page. Down at the bottom is a meter showing how tight the power system is on capacity right now. When that gets into the yellow ranges, clothes dryers and air conditioners need to reduce their power consumption, and if it goes into the red (which is rare), they need to shut down. Electricity rates should go up when the power situation is in the yellow and red.
It's not "your" printer. You don't own the software in the printer, or the driver, or the service that handles spamming you. You're just licensing that. You're renting a printing service, and the landlord controls what you can do with the printer.
Read your EULA.
Good idea. I'm impressed that they were able to cram Occam into an Atmel ATMega. Occam syntax is rather clunky by modern standards, but it gets the job done. It has a sane concurrency model and is safer than C.
It's a cute idea, but probably works about as well as self-diagnosis through Google.
More useful would be portable, automatic test equipment for blood samples. This is already available (PowerPoint file, will open with Open Office Impress) in a small desktop machine. It's marketed, though, as a device which allows US medical practitioners to do blood tests in their offices while charging insurance companies the usual price an outside lab would charge. These things need to come down to the cost of a high-end smartphone, be deployed in pharmacies, and operated by people at the nurse/practitioner, paramedic, or pharmacist level.
I had something like this happen a few years back. I have a domain in ".com" which is the same as the "co.uk" domain of a boarding school in England. Occasionally I'd get misaddressed mail. (This was back when you could use a catchall address for a domain without being overwhelmed by spam.) Once I got a message with the subject "I am going to kill you tonight". After checking the headers, it was clear that it was from someone at the school, not a death threat aimed at me. (Sent from.co.uk, addressed to same second level domain in.com.) Called up the school in England and reached someone in authority. 8 hour time difference; middle of the night there, someone had to be awakened. Turned out it was a 12-year old kid sending a dumb email to one of the other kids. He was disciplined by the school.
Personally I prefer working with ATmega's directly rather than with Arduino, but... if you want to futz around and LEARN, Arduino is a good place for it.
Yes. ATMega boards with small LCD displays are available, and Atmel's free AVR Studio is a reasonable IDE, with C and C++. If you already know how to program and don't want to join the Arduno cult, it's a reasonable way to get things done. There's a wide range of ATMega parts with different combinations of RAM, Flash, and I/O devices. AVR Studio supports all of them; the Arduno support is more limited.
A few years ago, there was a modernistic little tea shop in Palo Alto which not only had free WiFi, but electric outlets at every table. So the place was full of people with laptops. It was very quiet. Nobody talked.
They didn't buy much, though. The woman behind the counter had so little to do that she was usually reading
(a book, not a screen, typically some 19th century classic; she was a philosophy major.)
The place lasted about six months. Then it went over to being a coffee bar. That didn't work either.
Now it's a yogurt place, with few tables and no available power outlets.
The IEEE points out that, at present, only about 1/3 of electrical engineers have electrical engineering jobs.
They also point out that in 1970, electrical engineers and doctors made about the same amount of money.
Source the parts better. It sounds like they have pulled this diode from a display projector,
Yes, they admit they did that. So they just have a prototype.
There's no big secret about the laser diode. It's a Nichia NDB7352. Any legit company can order those things in bulk from Nichia in Tokyo.
No US distributor, including Nichia America, stocks them. WickedLasers probably doesn't buy enough of them to place an order with Nichia.
Industrial civilization has a big problem ahead.
It's been 50 years since the last new energy source was invented. (Atomic power and solar cells are more than 50 years old.) And we're running out.
Wind power seems promising, but the available sites are limited. There are four good onshore wind sites in California (Pacheco Pass, Altamont Pass, Mojave, and Montezuma Hills.) Each already has a big wind farm. There are plenty of good sites in the flyover states (from the Texas panhandle north to Canada) but it's hard to get power from there to someplace useful. Anyway, wind is too intermittent to be useful for more than 25-30% of power, tops.
All the good hydroelectric sites were developed by 1940 or so. And the big reservoirs behind the dams are silting up.
Nuclear power works, but all designs other than classic pressurized-water and boiling reactors have been flops. Gas-cooled reactors, pelletized-bed reactors, and thorium reactors require more complex components on the radioactive side of the system, and for each of those types, the prototypes have had serious problems.
Solar cells are useful, at least in the daytime. But they're expensive to make and take up a lot of land.
You only get about 50 watts per square meter over 24 hours, or about 1KWh per day.
Solar power satellites are a fantasy. We have no way of putting that much mass in orbit at a profit.
Tidal power is a niche technology. There are about ten good sites in the world.
Natural gas and coal are good for most of the next century, but they're not unlimited. Coal reserves in the US turn out to have been overestimated - the 200 year availability figure assumed that any coal in the ground could be strip-mined, even if there was a city or farmland on top of it. (Big issue in Illinois - the same land that's being used for farming has coal underneath.)
Ethanol from corn is a tax-funded boondoggle. Ethanol from cellulose, maybe, if the biotech people come through with bacteria able to convert cellulose into something more useful at low cost.
The problems with oil are well known.
Energy efficiency improvement is useful, but once you get past the first 25% or so, it gets hard. The more efficient equipment may in the end a lose.
And that's where we are, not much better off than 50 years ago, but closer to the end of nonrenewable resources.
Read the original article. Gou, the CEO of Foxconn, talked at their annual meeting about moving some production to Taiwan, Vietnam, and India. It's not clear that they even intend to reduce their head count in China; that's a speculation by Oriental Daily. Foxconn has been growing rapidly, and they have too many people at one location. (Managing really huge plants is historically a headache. The maximum optimal plant size seems to be around 3,000, from modern US experience. All the economies of scale have been achieved by then. China is at an earlier stage of automation, though. The US at one time had single steel plants that employed 8,000 people with shovels. )
Why should there be "billion dollar open source companies?" Most have nothing that valuable to sell. Red Hat is a packager; they don't implement much.
Remember VA Linux, the people behind Slashdot? Biggest first-day runup in the history of the NASDAQ. Where are they now? Down from 233 to 1.33.
Craigslist is the price leader in local ads, and they're a small company. Jimbo Wales thought Wikia was going to be a big deal (he wanted a private jet, like the Google people), but Wikia turned out to be merely a free hosting service for fancruft (the Star [Wars|Gate|Trek|Craft} wikis, etc.).
The USPTO has had their patent files on line for years. They're at the USPTO site. There's a reasonable search engine. Not only are the patents themselves there, but the whole "file wrapper" history information is available. However, it's rate limited to about 1000 patents per day per user.
What Google is doing, apparently, is making the entire database available for bulk download. The USPTO sells that database, and the other patent-indexing services buy it. But you have to get it on Digital Linear Tapes (DLT) and it costs several thousand dollars, because of the size of the data set. Google is apparently willing to put that data out for full download in exchange for getting their copy for free. And the USPTO gets out of the tape copying business.
They need to do some underground mapping and find other voids. That's routinely done for oil exploration, by using a "thumper truck" to pound on the ground while arrays of microphones listen. Suitable number-crunching yields an 3D underground map. Then they'll at least know where future problems are expected. That's not too expensive.
Fixing the problem is tougher. Drilling holes into voids and pumping in cement might work, but that's expensive, and it's going to divert water flows to some other area.
It's sad. I was at the Computer Museum in Mountain View a few years ago, where they had a Cray-I in a corner of the lobby, just sitting there used as a bench. It's not even labeled; some visitors think it's just furniture.
We need a general legal prohibition against implicit consent. We need to legislate that any contract provision which allows one party to modify the terms of the contract without the explicit consent of the other party is null and void. Consent cannot be given merely by continuing to use the service or pay for it.
If companies want to change terms, they should have to get explicit consent, given as a transaction separate from ordinary use of the product or service. This should be an FTC rule.
The effect is that when a company changes something, some of the customers are going to move to a competitor. That's in line with the basic principles of capitalism.
Slashdot Mirror
However, I will admit that this is certainly beyond the capabilities of most people due to the lack test equipment that is needed to even test the parts found at the swap meet.
True. Although it is easier to get gigahertz test gear than it used to be, the typical 'scope won't go there.
Link to actual project at General Electric, including access to the Edison audio.
I'm impressed with what this guy found at a hamfest. We don't see much microwave gear in Silicon Valley surplus any more. eBay, though, has a decent selection of microwave horns, low noise amplifiers, mixers, and waveguide. It looks like anybody could get the necessary parts in small quantity. New, though, those parts are expensive, so building low-cost robot vision systems this way is hard.
Also, when your "garage machine shop" has a Bridgeport milling machine, you're way above the usual home shop level. Still, if there's a TechShop in your town, you can get access to such machines.
A big problem working in this area, even if you know what you're doing, is that the test gear you need costs more than the thing you're making. Reading the design notes, some of which are on Air Force Research Lab stationery, indicate that the hamfest parts were tested and characterized using reasonably good test gear. And this was an MIT student, with access to MIT labs.
I ran into that building a small LIDAR in the early 1990s. The parts cost wasn't too bad, but I needed access to about $20K in test gear to debug the thing.
It's a real problem for Zuckerman. He's previously made fund-raising trips to Dubai. That's over. The UAE has blasphemy laws, which they enforce. The UAE also has an extradition treaty with Pakistan, but not with the United States. So he can no longer visit Dubai, and is unlikely to get funding from any source in the Arab world. He can't even fly Emirates Air.
Chess has finally been solved to the point that there's now unbeatable AIs available to the average user (assuming it gets to move first)...
No, checkers has been solved to that point. The solution is available online. Perfect play leads to a draw.
Computer chess is merely at the point that if you haven't been on the cover of Chess Life, you're going to get trounced. Even if you have, you're going to lose more than you win. The current situation is that Deep Rybka 2010 has an ELO rating around 3150. That's running on a 4-core AMD-64 desktop machine. The all-time human record is 2851, which Garry Kasparov had in 1999-2000.
The Man Whose Name Wouldn't Fit, published in 1968.
I'm just curious how long it will be before someone feels compelled to jailbreak their printer.
Printers already have 1) the code that prints the secret pattern of little yellow dots that identifies the printer to law enforcement, 2) the code that recognizes the pattern in European Union money and won't print that, and 3) the code that recognizes the proprietary ink cartridges.
With newer printers, you're pwned.
They're several years too late. The market for Wall Street "quants" has collapsed.
All this "smart meter" stuff has too much data flowing around, and too much of it is sent out of the home back to organizations which may make dubious use of it.
All that's really needed to level out peaks is to broadcast a few bits of information per hour to interested power-using devices. Here's the California Independent System Operator status page. Down at the bottom is a meter showing how tight the power system is on capacity right now. When that gets into the yellow ranges, clothes dryers and air conditioners need to reduce their power consumption, and if it goes into the red (which is rare), they need to shut down. Electricity rates should go up when the power situation is in the yellow and red.
It's not "your" printer. You don't own the software in the printer, or the driver, or the service that handles spamming you. You're just licensing that. You're renting a printing service, and the landlord controls what you can do with the printer. Read your EULA.
Good idea. I'm impressed that they were able to cram Occam into an Atmel ATMega. Occam syntax is rather clunky by modern standards, but it gets the job done. It has a sane concurrency model and is safer than C.
Next, Ada?
It's a cute idea, but probably works about as well as self-diagnosis through Google.
More useful would be portable, automatic test equipment for blood samples. This is already available (PowerPoint file, will open with Open Office Impress) in a small desktop machine. It's marketed, though, as a device which allows US medical practitioners to do blood tests in their offices while charging insurance companies the usual price an outside lab would charge. These things need to come down to the cost of a high-end smartphone, be deployed in pharmacies, and operated by people at the nurse/practitioner, paramedic, or pharmacist level.
Looks like a false alarm. Later report: "A 19-year-old man who was arrested by armed police at a Merseyside school has been released on bail. ". "Merseyside Police said that their inquiries were continuing into the man, who had imitation firearms and a computer seized from his home. The alert had been raised after a threat with a picture of a gun was posted on a social networking website."
I had something like this happen a few years back. I have a domain in ".com" which is the same as the "co.uk" domain of a boarding school in England. Occasionally I'd get misaddressed mail. (This was back when you could use a catchall address for a domain without being overwhelmed by spam.) Once I got a message with the subject "I am going to kill you tonight". After checking the headers, it was clear that it was from someone at the school, not a death threat aimed at me. (Sent from .co.uk, addressed to same second level domain in .com.) Called up the school in England and reached someone in authority. 8 hour time difference; middle of the night there, someone had to be awakened. Turned out it was a 12-year old kid sending a dumb email to one of the other kids. He was disciplined by the school.
Today, they'd send in a SWAT team.
Personally I prefer working with ATmega's directly rather than with Arduino, but ... if you want to futz around and LEARN, Arduino is a good place for it.
Yes. ATMega boards with small LCD displays are available, and Atmel's free AVR Studio is a reasonable IDE, with C and C++. If you already know how to program and don't want to join the Arduno cult, it's a reasonable way to get things done. There's a wide range of ATMega parts with different combinations of RAM, Flash, and I/O devices. AVR Studio supports all of them; the Arduno support is more limited.
A few years ago, there was a modernistic little tea shop in Palo Alto which not only had free WiFi, but electric outlets at every table. So the place was full of people with laptops. It was very quiet. Nobody talked.
They didn't buy much, though. The woman behind the counter had so little to do that she was usually reading (a book, not a screen, typically some 19th century classic; she was a philosophy major.) The place lasted about six months. Then it went over to being a coffee bar. That didn't work either. Now it's a yogurt place, with few tables and no available power outlets.
The IEEE points out that, at present, only about 1/3 of electrical engineers have electrical engineering jobs. They also point out that in 1970, electrical engineers and doctors made about the same amount of money.
Lawyers, though, are starting to get hit. Outsourcing of legal work is now available.
Source the parts better. It sounds like they have pulled this diode from a display projector,
Yes, they admit they did that. So they just have a prototype.
There's no big secret about the laser diode. It's a Nichia NDB7352. Any legit company can order those things in bulk from Nichia in Tokyo. No US distributor, including Nichia America, stocks them. WickedLasers probably doesn't buy enough of them to place an order with Nichia.
Industrial civilization has a big problem ahead. It's been 50 years since the last new energy source was invented. (Atomic power and solar cells are more than 50 years old.) And we're running out.
Wind power seems promising, but the available sites are limited. There are four good onshore wind sites in California (Pacheco Pass, Altamont Pass, Mojave, and Montezuma Hills.) Each already has a big wind farm. There are plenty of good sites in the flyover states (from the Texas panhandle north to Canada) but it's hard to get power from there to someplace useful. Anyway, wind is too intermittent to be useful for more than 25-30% of power, tops.
All the good hydroelectric sites were developed by 1940 or so. And the big reservoirs behind the dams are silting up.
Nuclear power works, but all designs other than classic pressurized-water and boiling reactors have been flops. Gas-cooled reactors, pelletized-bed reactors, and thorium reactors require more complex components on the radioactive side of the system, and for each of those types, the prototypes have had serious problems.
Solar cells are useful, at least in the daytime. But they're expensive to make and take up a lot of land. You only get about 50 watts per square meter over 24 hours, or about 1KWh per day.
Solar power satellites are a fantasy. We have no way of putting that much mass in orbit at a profit.
Tidal power is a niche technology. There are about ten good sites in the world.
Natural gas and coal are good for most of the next century, but they're not unlimited. Coal reserves in the US turn out to have been overestimated - the 200 year availability figure assumed that any coal in the ground could be strip-mined, even if there was a city or farmland on top of it. (Big issue in Illinois - the same land that's being used for farming has coal underneath.)
Ethanol from corn is a tax-funded boondoggle. Ethanol from cellulose, maybe, if the biotech people come through with bacteria able to convert cellulose into something more useful at low cost.
The problems with oil are well known.
Energy efficiency improvement is useful, but once you get past the first 25% or so, it gets hard. The more efficient equipment may in the end a lose.
And that's where we are, not much better off than 50 years ago, but closer to the end of nonrenewable resources.
Read the original article. Gou, the CEO of Foxconn, talked at their annual meeting about moving some production to Taiwan, Vietnam, and India. It's not clear that they even intend to reduce their head count in China; that's a speculation by Oriental Daily. Foxconn has been growing rapidly, and they have too many people at one location. (Managing really huge plants is historically a headache. The maximum optimal plant size seems to be around 3,000, from modern US experience. All the economies of scale have been achieved by then. China is at an earlier stage of automation, though. The US at one time had single steel plants that employed 8,000 people with shovels. )
Why should there be "billion dollar open source companies?" Most have nothing that valuable to sell. Red Hat is a packager; they don't implement much.
Remember VA Linux, the people behind Slashdot? Biggest first-day runup in the history of the NASDAQ. Where are they now? Down from 233 to 1.33.
Craigslist is the price leader in local ads, and they're a small company. Jimbo Wales thought Wikia was going to be a big deal (he wanted a private jet, like the Google people), but Wikia turned out to be merely a free hosting service for fancruft (the Star [Wars|Gate|Trek|Craft} wikis, etc.).
The USPTO has had their patent files on line for years. They're at the USPTO site. There's a reasonable search engine. Not only are the patents themselves there, but the whole "file wrapper" history information is available. However, it's rate limited to about 1000 patents per day per user.
What Google is doing, apparently, is making the entire database available for bulk download. The USPTO sells that database, and the other patent-indexing services buy it. But you have to get it on Digital Linear Tapes (DLT) and it costs several thousand dollars, because of the size of the data set. Google is apparently willing to put that data out for full download in exchange for getting their copy for free. And the USPTO gets out of the tape copying business.
So how much is a ring?
They need to do some underground mapping and find other voids. That's routinely done for oil exploration, by using a "thumper truck" to pound on the ground while arrays of microphones listen. Suitable number-crunching yields an 3D underground map. Then they'll at least know where future problems are expected. That's not too expensive.
Fixing the problem is tougher. Drilling holes into voids and pumping in cement might work, but that's expensive, and it's going to divert water flows to some other area.
It's sad. I was at the Computer Museum in Mountain View a few years ago, where they had a Cray-I in a corner of the lobby, just sitting there used as a bench. It's not even labeled; some visitors think it's just furniture.
We need a general legal prohibition against implicit consent. We need to legislate that any contract provision which allows one party to modify the terms of the contract without the explicit consent of the other party is null and void. Consent cannot be given merely by continuing to use the service or pay for it. If companies want to change terms, they should have to get explicit consent, given as a transaction separate from ordinary use of the product or service. This should be an FTC rule.
The effect is that when a company changes something, some of the customers are going to move to a competitor. That's in line with the basic principles of capitalism.