Just about every "Bitcoin exchange" has had some huge problem. Either they take the money and run, or they get broken into and lose the money.
Tradehill sounded like the most legitimate of the exchanges. Then, in February 2012, Tradehill shut down with no notice, after a big chargeback from Dwolla. They did refund the money, though.
Bitcoinica collapsed and lost customer funds. There was the Bruce Wagner fiasco. Below that level, there were about a dozen other "exchanges" and "online wallets" that lost customer funds.
Mt. Gox is almost the only exchange that hasn't yet lost customer funds or collapsed. Even there, one wonders. They're vague about who's really behind the organization. Mt. Gox started as "Magic, the Gathering Online Exchange."
Meanwhile, ARM tablets on Amazon start around $60, and ones with decent reviews start around $80. Most run the open source version of Android. Google's Nexus 7 costs $199, and that's the price point Microsoft will probably have to match, if not beat.
From a marketing perspective, once the price of something drops below $100, sales go way up.
Much employee-owned stock couldn't be sold until the first lockup period ended. Which it just did. So, given Facebook's declining stock price, it's time to cash out. Of course they're quitting. Facebook is profitable, but the stock is overpriced by an order of magnitude or so.
Lockups are far shorter than they used to be. When I cashed out of Autodesk in the 1980s, insiders had a 2-year lockup on restricted stock. And you had to pay taxes when you exercised an option, even though you couldn't sell for another two years. That was before "deregulation", and kept insiders from cashing out before the company tanked. Now it's 90 to 277 days.
This encourages hyping the stock, taking the money, and running.
Google didn't build on-campus but they signed up with Tech Shop a while back.
Google does have an on-campus shop for employees. They're not doing much at TechShop. They've done team-building events at the Menlo Park TechShop, and downloaded and cut out a model of a bridge on the plasma cutter. One or two Google employees go there now and then. But it's not a big thing.
Tablet computers are becoming a commodity.
A 7" tablet from China is only $70. On Amazon, you can now get Android tablets from $60.
Since the Allwinner ARM system on a chip came out for $7, with no US intellectual property to run up the price, the compute power in low-end tablets has been quite impressive. Tablet computers are going to be something you buy in a blister pack at the convenience store.
How will Apple, with all their expensive stores on expensive real estate, and a business built on huge markups, deal with that? Their pricing is around $400, over five times the price of the competition. They can't maintain that margin.
There's a market for luxury items. The CEO of Rolex says "We are not in the watch business, we are in the luxury business. The volumes are small. Apple is too big a company to take that route. Apple may have to try coming out with lower-priced lines to compete.
Look it's doesn't matter how many petitions you stand up. Basically the folks that have the authority and power to control the people, will. Common folk are only here to support the rich and powerful by way of their taxes. Nothing else matters. You're either part of the good-old-boy network, or you're nobody. It's always been this way; for every country; for every regime; for every global power, since time began.
That wasn't true of the US from WWII to about 1960. Truman and Eisenhower were modest people. Truman ran a hat store. Eisenhower was a night supervisor at a creamery before he got into West Point. That period was probably the most successful in American history.
Here, from GitHub, is the short presentation. This is very impressive. It finds parallelism automatically, at least for simple cases. Over 50x performance improvement on matrix multiply and naive Fourier transform (not FFT), both of which have very simple inner loops. Not clear how it does on less obvious problems.
There are "Trash Hunter boats that could pick up pumice, but they're not intended for remote open-ocean operations. To collect this stuff, it would take booms and ocean-going tugs or fishing boats to concentrate the floating pumice, a collection vessel to pull it out of the water and screen it, and a bulk freighter to haul it to some customer. It's like cleaning up an oil spill, except that it's a solid. It might be desirable to do this if the mess drifts to a populated area.
Over time, wave action breaks the stuff up, opens the gas pockets that make it float, and it sinks. This takes about a year, so it's not a long term problem. It happens now and then. Known events off Tonga in 1964 and 2002 have been studied. Long-term impact is low; it's hard to tell, a few years later, that it ever happened.
The technology industry is a little too obsessed with 'sending little messages to each other and having fun on a social network'
Regardless of who said it, that's a good point. Computing has become a branch of the advertising industry. As computing has become cheaper, the applications have become more banal.
This is a real problem for society. Advertising mostly moves demand around. With most people in the US spending all their disposable income, it doesn't create new demand. What it does do is drive up product prices. There are many products, from movies to prescription drugs, where the marketing cost exceeds the manufacturing cost.
As an actuator, Nitinol can produce significant power in small package, but it's a very inefficient device. The metal will change crystal structure when heated, and return to the original shape when cooled. Heating is usually accomplished by running electricity through the Nitinol wire. Most of the energy goes into waste heat; only a small fraction comes out of the actuator as useful work.
So a battery-powered earthworm isn't likely. As a cabled device, it has potential. A great application would be short run cable-laying for fibre optics. A machine that could get a fibre optic cable underground from street to house without digging up sidewalks and lawns would be very useful.
I have an MSCS from Stanford, but it's from 1985, when the logicians and expert systems guys were running things. So I have lots of number theory, combinatorics, automata theory, and mathematical logic. I even took "Epistemological Problems in Artificial Intelligence" from John McCarthy.
So what did I end up needing? Tensor calculus. I realized that expert systems AI was stuck. The future of AI capable of dealing with the real world seemed to be in nonlinear control theory. Which is all calculus and statistics. I struggled with that, and got legged running over rough terrain figured out and patented. But this was 1994, and the simulators sucked, and I couldn't get any further without better simulators. So I spent a few years beating on that problem, and produced the first simulator that could do a ragdoll falling downstairs.
By 1997, I had that solved, but it was kind of slow. A 200MHz Pentium Pro just wasn't enough engine to get it up to real time, and that was the top of the line in CPUs back then. By then I was burnt out on the problem, and it wasn't making much money, so I sold the technology off to Havok and went on to other things.
I didn't see that what was needed was to couple nonlinear control theory to Bayesian statistics. That's what makes all those quadrotors zip around so precisely. Modern statistics barely existed when I was in school. Now it drives everything from finance to speech recognition to advertising, so it gets worked on and people study it. Nonlinear control alone never had that big a market, so the field didn't get enough attention to move it forward.
So I needed more math, and different math, than I got in school.
Just before Webvan tanked, they were playing around with a similar concept. In the Webvan system, the lockers were refrigerated, since they were delivering food. The concept was to have locker rooms in large apartment buildings in big cities, so you could order food and have it ready for you when you got home. Great idea for NYC and London, where people try to carry groceries on the subway or are stuck shopping at overpriced local shops with small selections.
Webvan is back. It's now owned by Amazon. They don't do perishables, and delivery takes 2-3 days, but you can order about 45,000 food-related products. Amazon plans to grow that business.
This could wipe many more retailers off the face of the earth. If the delivery density is high enough, delivery is cheaper that driving a 2-ton SUV to a mall for 20 pounds of groceries.
This has been proposed before. It's already obsolete.
The Uniform Resource Name idea was supposed to do this. So was the "Semantic Web". In practice, there are many edge caching systems already, Akamai being the biggest provider.
Most networking congestion problems today are at the edges, where they should be, not at the core. Bulk bandwidth is cheap.
The concept is obsolete because so much content is now "personalized". You can't cache a Facebook page or a Google search result. Every serve of the same URL produces different output. Video can be cached or multicast only if the source of the video doesn't object. Many video content sources would consider it a copyright violation. Especially if it breaks ad personalization.
As for running out of bandwidth, we're well on our way to enough capacity to stream HDTV to everybody on the planet simultaneously. Beyond that, it's hard to usefully use more bandwidth. Wireless spectrum space is a problem, but caching won't help there.
The sheer amount of infrastructure that's been deployed merely so that people can watch TV over the Internet is awe-inspiring. Arguably it could have been done more efficiently, but if it had been, it would have been worse. Various schemes were proposed by the cable TV industry over the last two decades, most of which were ways to do pay-per-view at lower cost to the cable company. With those schemes, the only content you could watch was sold by the cable company. We're lucky to have escaped that fate.
Barracuda Networks doesn't even seem to have published a paper. (The article linked in the Slashdot article is a scraper site for press releases.) The Barracuda press release points to an "infographic" and a blog posting which, as their only outside source, links to a black hat site.
Barracuda doesn't seem to have discovered the extent of the social spamming ecosystem. We identified at least 6 levels:
Advertising agencies.
SEO firms. ("Google Places Guaranteed")
Fake review, "like", "+1", and "retweet" generators. ("Buy Facebook Fans with
us today and watch your popularity boom.")
Fake account generators, both automated and outsourced to low-wage countries. ("Bulk Accounts is the largest mass account generator out there....Gmail, Myspace, Youtube, Facebook, Twitter, Hotmail and much more...")
Fake IP address proxies and fake phone numbers ("Premium Private Proxies", "Top Quality CL Phone Numbers used to
create Craigslist PVAs")
Botnet operators providing proxies on compromised machines. Now we're down at the organized crime level.
This structure insulates the legitimate businesses who use ad agencies from the criminal activity at the bottom.
Except for the botnet operators, everybody in that ecosystem has some kind of web presence, although towards the bottom, they usually have only Skype and Gmail accounts as contacts. I'm not going to link to them here, but our paper gives actual names.
In Heinlein's Future History series from the 1950s, there is a time line chart. This chart shows a "false dawn" in space travel - initial success around 1970, then a long hiatus.
In Heinlein's "The Man who Sold The Moon", the problem is made clear - fuel. A chemically powered rocket can just barely make it to the moon, with severe weight restrictions. Nuclear rockets are too dangerous. And so, the first lunar landing is a publicity stunt.
Heinlein could do the math. Space travel with chemical rockets is just barely feasible and hugely expensive. Nuclear rocket engines were built and successfully tested in the 1950s, but are too dangerous to use. Fusion isn't even close to working. So we're stuck.
Diode-type CMOS imagers for terahertz radiation have been built. Those convert terahertz radiation into DC, which can then be amplified by standard techniques. But diodes don't have gain. That's why the original article emphasizes that this new device has gain.
There are terahertz lasers, waveguides, antennas, and other components that work up there. The situation is much like radar during WWII; there were a few components that could do specific things at radar frequencies (then 60MHz to 1.2GHz), but general electronics wasn't there yet. Most of the electronics in radars of that period ran at far lower speeds. They still worked.
Firefox suffers from an antiquated code base. It's single-thread, and the project to make it multi-thread failed.
There are two interpretive systems inside - Javascript and XUL - and they aren't on good speaking terms. There are two plug-in systems, "classic" and "Jetpack", and the teams for those sometimes don't seem to be on speaking terms. The number of open bugs keeps creeping up, and much bug-closing is "developer in denial", not an actual fix. Startup is slow, and, at times, shutdown is even slower.
Then came the frantic release cycle, which didn't help.
Why would you have an account on Google? Their search works fine without it. Their video streaming works fine without it. If your ISP has an IMAP server and maybe some form of webmail for emergencies, that takes care of mail. None of their other services are worth much.
The one failing of the Google feedback system, IMO, is that it lacks feedback.
Which means they're wasting your time. As a general policy, I never use "feedback" systems that don't generate a response, a public posting, or a ticket number.
Here's the National Advanced Driving Simulator, which is in Iowa. This not only has a Stewart platform, the Stewart Platform is mounted on an X-Y table about 60 feet square. Toyota has an even bigger one with over 100 feet of linear travel.
The need for huge linear travel comes from the need to simulate the feeling of a hard stop. To some extent, deceleration can be simulated with tilt. But at the end of a stop, deceleration suddenly ceases without a change in attitude. You can't simulate that with a Stewart platform. If you want to test people's behavior during hard braking, you need a huge simulator.
The graphics are scruffy and the music is annoying first thing in the morning, but it's very impressive.
True. That was made on a Mac IIci, and the audio is 8-bit. It took hours to run a simulation like that back then. I spent a few years after that working on simulation technology.
Simulating real legged machines with realistic simulation of ground contact is still a CPU hog today. Video games mostly fake their contact physics.
The new mail client has the Metro look and feel. And it is providing users with more granular control over which ads they see and where they see them.'"
Slashdot Mirror
Just about every "Bitcoin exchange" has had some huge problem. Either they take the money and run, or they get broken into and lose the money.
Tradehill sounded like the most legitimate of the exchanges. Then, in February 2012, Tradehill shut down with no notice, after a big chargeback from Dwolla. They did refund the money, though.
Bitcoinica collapsed and lost customer funds. There was the Bruce Wagner fiasco. Below that level, there were about a dozen other "exchanges" and "online wallets" that lost customer funds.
Mt. Gox is almost the only exchange that hasn't yet lost customer funds or collapsed. Even there, one wonders. They're vague about who's really behind the organization. Mt. Gox started as "Magic, the Gathering Online Exchange."
Acer is begging Microsoft not to price the Microsoft Surface tablet at $199. Acer would like to see it priced around $499-$599. Otherwise it will "seriously impact the existing PC ecosystem".
Meanwhile, ARM tablets on Amazon start around $60, and ones with decent reviews start around $80. Most run the open source version of Android. Google's Nexus 7 costs $199, and that's the price point Microsoft will probably have to match, if not beat.
From a marketing perspective, once the price of something drops below $100, sales go way up.
Much employee-owned stock couldn't be sold until the first lockup period ended. Which it just did. So, given Facebook's declining stock price, it's time to cash out. Of course they're quitting. Facebook is profitable, but the stock is overpriced by an order of magnitude or so.
Lockups are far shorter than they used to be. When I cashed out of Autodesk in the 1980s, insiders had a 2-year lockup on restricted stock. And you had to pay taxes when you exercised an option, even though you couldn't sell for another two years. That was before "deregulation", and kept insiders from cashing out before the company tanked. Now it's 90 to 277 days. This encourages hyping the stock, taking the money, and running.
Google didn't build on-campus but they signed up with Tech Shop a while back.
Google does have an on-campus shop for employees. They're not doing much at TechShop. They've done team-building events at the Menlo Park TechShop, and downloaded and cut out a model of a bridge on the plasma cutter. One or two Google employees go there now and then. But it's not a big thing.
Tablet computers are becoming a commodity. A 7" tablet from China is only $70. On Amazon, you can now get Android tablets from $60. Since the Allwinner ARM system on a chip came out for $7, with no US intellectual property to run up the price, the compute power in low-end tablets has been quite impressive. Tablet computers are going to be something you buy in a blister pack at the convenience store.
How will Apple, with all their expensive stores on expensive real estate, and a business built on huge markups, deal with that? Their pricing is around $400, over five times the price of the competition. They can't maintain that margin.
There's a market for luxury items. The CEO of Rolex says "We are not in the watch business, we are in the luxury business. The volumes are small. Apple is too big a company to take that route. Apple may have to try coming out with lower-priced lines to compete.
Look it's doesn't matter how many petitions you stand up. Basically the folks that have the authority and power to control the people, will. Common folk are only here to support the rich and powerful by way of their taxes. Nothing else matters. You're either part of the good-old-boy network, or you're nobody. It's always been this way; for every country; for every regime; for every global power, since time began.
That wasn't true of the US from WWII to about 1960. Truman and Eisenhower were modest people. Truman ran a hat store. Eisenhower was a night supervisor at a creamery before he got into West Point. That period was probably the most successful in American history.
Here, from GitHub, is the short presentation. This is very impressive. It finds parallelism automatically, at least for simple cases. Over 50x performance improvement on matrix multiply and naive Fourier transform (not FFT), both of which have very simple inner loops. Not clear how it does on less obvious problems.
Not sure where you got that price from.
USGS Minerals Industry Summary - Pumice. That's the bulk price.
There are "Trash Hunter boats that could pick up pumice, but they're not intended for remote open-ocean operations. To collect this stuff, it would take booms and ocean-going tugs or fishing boats to concentrate the floating pumice, a collection vessel to pull it out of the water and screen it, and a bulk freighter to haul it to some customer. It's like cleaning up an oil spill, except that it's a solid. It might be desirable to do this if the mess drifts to a populated area.
Over time, wave action breaks the stuff up, opens the gas pockets that make it float, and it sinks. This takes about a year, so it's not a long term problem. It happens now and then. Known events off Tonga in 1964 and 2002 have been studied. Long-term impact is low; it's hard to tell, a few years later, that it ever happened.
It could be scooped up and sold, but at $25 a ton, it's not worth it.
The technology industry is a little too obsessed with 'sending little messages to each other and having fun on a social network'
Regardless of who said it, that's a good point. Computing has become a branch of the advertising industry. As computing has become cheaper, the applications have become more banal.
This is a real problem for society. Advertising mostly moves demand around. With most people in the US spending all their disposable income, it doesn't create new demand. What it does do is drive up product prices. There are many products, from movies to prescription drugs, where the marketing cost exceeds the manufacturing cost.
Can we still afford this?
This "earthworm" device uses Nitinol shape-memory alloy as an actuator. That's been tried many times before, going back to the 1980s.
As an actuator, Nitinol can produce significant power in small package, but it's a very inefficient device. The metal will change crystal structure when heated, and return to the original shape when cooled. Heating is usually accomplished by running electricity through the Nitinol wire. Most of the energy goes into waste heat; only a small fraction comes out of the actuator as useful work.
So a battery-powered earthworm isn't likely. As a cabled device, it has potential. A great application would be short run cable-laying for fibre optics. A machine that could get a fibre optic cable underground from street to house without digging up sidewalks and lawns would be very useful.
It's hard to imagine being found at-fault when you are stopped and rear-ended.
Especially when the self-driving car has full video, lidar, and radar coverage of the entire event. And really good lawyers.
I have an MSCS from Stanford, but it's from 1985, when the logicians and expert systems guys were running things. So I have lots of number theory, combinatorics, automata theory, and mathematical logic. I even took "Epistemological Problems in Artificial Intelligence" from John McCarthy.
So what did I end up needing? Tensor calculus. I realized that expert systems AI was stuck. The future of AI capable of dealing with the real world seemed to be in nonlinear control theory. Which is all calculus and statistics. I struggled with that, and got legged running over rough terrain figured out and patented. But this was 1994, and the simulators sucked, and I couldn't get any further without better simulators. So I spent a few years beating on that problem, and produced the first simulator that could do a ragdoll falling downstairs.
By 1997, I had that solved, but it was kind of slow. A 200MHz Pentium Pro just wasn't enough engine to get it up to real time, and that was the top of the line in CPUs back then. By then I was burnt out on the problem, and it wasn't making much money, so I sold the technology off to Havok and went on to other things.
I didn't see that what was needed was to couple nonlinear control theory to Bayesian statistics. That's what makes all those quadrotors zip around so precisely. Modern statistics barely existed when I was in school. Now it drives everything from finance to speech recognition to advertising, so it gets worked on and people study it. Nonlinear control alone never had that big a market, so the field didn't get enough attention to move it forward.
So I needed more math, and different math, than I got in school.
Microsoft got into search as a defensive move against Google. They lost money on it.
Google got into online documents and spreadsheets as a competitve move against Microsoft's moneymaker, Office. That didn't make Google money.
Google got into phone software as a defensive move against Apple. That hasn't made money.
News Corp bought Myspace as a defensive move against online. That was a disaster.
Time/Warner merged with AOL to get into "the Internet". That was a disaster and they broke up.
Google got into social as a defensive move against Facebook. So far, that hasn't worked out too well.
Facebook got into photo sharing as a defensive move against Yahoo's Flickr. That didn't work.
Just before Webvan tanked, they were playing around with a similar concept. In the Webvan system, the lockers were refrigerated, since they were delivering food. The concept was to have locker rooms in large apartment buildings in big cities, so you could order food and have it ready for you when you got home. Great idea for NYC and London, where people try to carry groceries on the subway or are stuck shopping at overpriced local shops with small selections.
Webvan is back. It's now owned by Amazon. They don't do perishables, and delivery takes 2-3 days, but you can order about 45,000 food-related products. Amazon plans to grow that business.
This could wipe many more retailers off the face of the earth. If the delivery density is high enough, delivery is cheaper that driving a 2-ton SUV to a mall for 20 pounds of groceries.
This has been proposed before. It's already obsolete.
The Uniform Resource Name idea was supposed to do this. So was the "Semantic Web". In practice, there are many edge caching systems already, Akamai being the biggest provider. Most networking congestion problems today are at the edges, where they should be, not at the core. Bulk bandwidth is cheap.
The concept is obsolete because so much content is now "personalized". You can't cache a Facebook page or a Google search result. Every serve of the same URL produces different output. Video can be cached or multicast only if the source of the video doesn't object. Many video content sources would consider it a copyright violation. Especially if it breaks ad personalization.
As for running out of bandwidth, we're well on our way to enough capacity to stream HDTV to everybody on the planet simultaneously. Beyond that, it's hard to usefully use more bandwidth. Wireless spectrum space is a problem, but caching won't help there.
The sheer amount of infrastructure that's been deployed merely so that people can watch TV over the Internet is awe-inspiring. Arguably it could have been done more efficiently, but if it had been, it would have been worse. Various schemes were proposed by the cable TV industry over the last two decades, most of which were ways to do pay-per-view at lower cost to the cable company. With those schemes, the only content you could watch was sold by the cable company. We're lucky to have escaped that fate.
Our paper from November 2011, "Social is bad for search, and search is bad for social", covered this last year.
Barracuda Networks doesn't even seem to have published a paper. (The article linked in the Slashdot article is a scraper site for press releases.) The Barracuda press release points to an "infographic" and a blog posting which, as their only outside source, links to a black hat site.
Barracuda doesn't seem to have discovered the extent of the social spamming ecosystem. We identified at least 6 levels:
This structure insulates the legitimate businesses who use ad agencies from the criminal activity at the bottom. Except for the botnet operators, everybody in that ecosystem has some kind of web presence, although towards the bottom, they usually have only Skype and Gmail accounts as contacts. I'm not going to link to them here, but our paper gives actual names.
In Heinlein's Future History series from the 1950s, there is a time line chart. This chart shows a "false dawn" in space travel - initial success around 1970, then a long hiatus.
In Heinlein's "The Man who Sold The Moon", the problem is made clear - fuel. A chemically powered rocket can just barely make it to the moon, with severe weight restrictions. Nuclear rockets are too dangerous. And so, the first lunar landing is a publicity stunt.
Heinlein could do the math. Space travel with chemical rockets is just barely feasible and hugely expensive. Nuclear rocket engines were built and successfully tested in the 1950s, but are too dangerous to use. Fusion isn't even close to working. So we're stuck.
Another terrible article summary.
In 2010, a solid-state device at 0.67THz was achieved. In 2012, that effort is up to 0.85 THz. Progress is slow, but continuing.
Diode-type CMOS imagers for terahertz radiation have been built. Those convert terahertz radiation into DC, which can then be amplified by standard techniques. But diodes don't have gain. That's why the original article emphasizes that this new device has gain.
There are terahertz lasers, waveguides, antennas, and other components that work up there. The situation is much like radar during WWII; there were a few components that could do specific things at radar frequencies (then 60MHz to 1.2GHz), but general electronics wasn't there yet. Most of the electronics in radars of that period ran at far lower speeds. They still worked.
Firefox suffers from an antiquated code base. It's single-thread, and the project to make it multi-thread failed. There are two interpretive systems inside - Javascript and XUL - and they aren't on good speaking terms. There are two plug-in systems, "classic" and "Jetpack", and the teams for those sometimes don't seem to be on speaking terms. The number of open bugs keeps creeping up, and much bug-closing is "developer in denial", not an actual fix. Startup is slow, and, at times, shutdown is even slower.
Then came the frantic release cycle, which didn't help.
Why would you have an account on Google? Their search works fine without it. Their video streaming works fine without it. If your ISP has an IMAP server and maybe some form of webmail for emergencies, that takes care of mail. None of their other services are worth much.
The one failing of the Google feedback system, IMO, is that it lacks feedback.
Which means they're wasting your time. As a general policy, I never use "feedback" systems that don't generate a response, a public posting, or a ticket number.
Here's the National Advanced Driving Simulator, which is in Iowa. This not only has a Stewart platform, the Stewart Platform is mounted on an X-Y table about 60 feet square. Toyota has an even bigger one with over 100 feet of linear travel.
The need for huge linear travel comes from the need to simulate the feeling of a hard stop. To some extent, deceleration can be simulated with tilt. But at the end of a stop, deceleration suddenly ceases without a change in attitude. You can't simulate that with a Stewart platform. If you want to test people's behavior during hard braking, you need a huge simulator.
The graphics are scruffy and the music is annoying first thing in the morning, but it's very impressive.
True. That was made on a Mac IIci, and the audio is 8-bit. It took hours to run a simulation like that back then. I spent a few years after that working on simulation technology.
Simulating real legged machines with realistic simulation of ground contact is still a CPU hog today. Video games mostly fake their contact physics.
The new mail client has the Metro look and feel. And it is providing users with more granular control over which ads they see and where they see them.'"
Ads? What do ads have to do with email?