I've worked on Mturk since it began in November of 2005 and I've earned $25,000 in all since then. I think the rate of pay is horrible but it's been fun trying to find ways to complete tasks more easily using tools such as GreaseMonkey and custom programming. Out and out robots aren't allowed but tools that reformat the screen for easier reading and clicking and things like that are permitted.
The pay was better when Mturk started but it's been a race to the bottom since then. Still, working from home when I want with almost no responsibility is so appealing that I still do it for the low pay. I can make maybe $5 to $10 per day and spend it on any stupid thing I want without cutting into my regular budget.
Reading TFA, it seems to me that all the safety features are proactive, i.e, they seek to prevent an accident. I doubt that zero-fatalities can be achieved that way. People will ignore the warnings, override the controls, etc. Rather, I think it would be better to concentrate on preventing fatalities when accidents occur using such techniques as enhanced air bags and reliable fire suppression systems. I'd be willing to pay more for those features.
I stumbled across the Khan lectures a couple of months ago and I think they're great. Many thanks to Salman Khan.
Another series of lectures that appeared on television about 30 years ago is a show called The Mechanical Universe. They're not as deep as some of the Khan lectures but a bit more polished (not that it matters a lot). The Mechanical Universe lectures are available online at: http://www.learner.org/resources/series42.html
Too bad there isn't more of this kind of material on television. Maybe one of the cable companies could buy the Khan lectures and make them available as an on-demand feature.
And what do you propose is a productive employment of capital? Reemployment in the casino scheme? Isn't that the most likely thing to happen in practice? Nobody is producing anything.
Your scissors thought experiment is an interesting example of non-physical faster-than-light travel. Along the same lines, here's something I've thought about from time to time and don't know if it makes sense:
Imagine a solid rod that reaches from NY City, NY to San Fransisco, CA. I want to send a signal from NY to CA so I push on the rod and move it one inch toward CA. An observer in CA is watching the rod and when he sees it move he knows I've sent my signal. My question is, how long does it take the signal to reach CA? Doesn't it seem on first thought to be instantaneous? If so, isn't it ftl signaling?
I'm probably overlooking something. Perhaps what is travelling along the bar is a compression wave at some finite speed. But that doesn't seem right because a solid rod, steel for example, doesn't compress an inch when you push on it.
Can anyone tell me what would really happen physically if this experiment was performed?
Not only do flash traders have faster access to data than others, they're given a commission discount by Goldman Sachs because of their volume of trading. If non-insiders tried to make money via arbitrage as the flash traders do, all of their profit would be negated by commissions. The game is definitely rigged.
"On Thursday afternoon, the Dow plunged 1,000 points within a few minutes, followed by an equally sudden recovery. We don’t know all the details about the drop, but it was almost certainly the result of computer or human error in a high-speed trading program.
Among the many arcane corners of the financial world highlighted by the Wall Street crisis, high-frequency trading — in which computers scan billions of bits of market data for trading opportunities that may exist for mere fractions of a second — has generated a surprising amount of discussion. Alongside the risk of expensive errors like what happened Thursday, critics say, these programs facilitate insider trading and overwhelm regulators’ access to critical information.
These are fair criticisms. Fortunately, they can also be easily addressed without undermining the positive role that high-frequency trading plays in the market.
Let’s start with the insider trading charge. Often, when an exchange operator receives an investor order and finds that another exchange has a better price, it will “flash” the order to a few select traders in its exchange a split second before sending it to market, giving those traders an opportunity to improve their price, too. When used properly, flashing ensures that investors trade at the best available prices.
But that hair’s breadth of time also gives high-frequency traders an opportunity to make a tidy profit off what amounts to insider information. How? Rather than improve their price, the recipient of a flash can go to the other exchange, buy up all the assets at better prices, and force the original investor to trade with them at an inferior price.
We don’t allow trading based on private knowledge of pending business deals or court rulings, and we shouldn’t allow it in high-frequency trading, either. But that doesn’t mean we should ban flashing all together. Instead, to deter abuse, anyone who gets a preview of a trade, whether by phone or flash, should be required to register with an exchange and keep records of every negotiation."
The review states that VNC isn't firewall friendly but, apparently, the reviewer isn't aware of the single click versions of VNC. These versions run without any changes to the host firewall since the connection is initiated by the host. The single click version of UltraVNC is available at: http://www.uvnc.com/addons/singleclick.html
I'm wondering what the risks are in using verbal checks (paperless ACH transfers). I pay my monthly electricity bill that way since my power company adds a "convenience charge" for using a credit card. As far as I know, the only thing needed for a verbal check is the account number and bank routing number. What's to stop anyone who knows the account number from issuing a verbal check to themself? The routing number for any bank is available online or by calling the bank. If I dispute a verbal check is the bank required to reverse the charge?
Got a CAT scanner sitting around? Stick the book under it and image each page as a thin slice.
I don't really know if this is technically feasible. How does ink react to xrays? How thin can a image slice be? How fast does a CAT scanner run? Maybe someone with CAT scanner knowledge can fill me in.
I know it's bad form to reply to my own post but I just found the following VirusTotal scan result for a rooted atapi.sys file that causes the blue screen:
My machines are running ok but I thought I'd look to see if KB977165 which is reported to cause the blue screens was on any of them. It was installed on one machine and not on another. So I thought I'd check a few more things that others say may be causing the problem.
It has been suggested that atapi.sys in the system32/drivers folder might be rootkitted by the update. I compared the SHA-1 hash of atapi.sys on both machines and they were the same (A719156E8AD67456556A02C34E762944234E7A44) so, apparently, update KB977165 didn't change that file in my case.
Some people are saying that atapi.sys is infected with a rootkit. I ran scanned the file through Jotti.com and it found nothing. I also scanned the file at VirusTotal.com and only eSafe reported a problem as follows: eSafe 7.0.17.0 2010.02.11 Win32.Rootkit
I think this is a false positive since I have identical copies of atapi.sys on both machines and both show the same result on eSafe.
It's possible, I suppose, that I was rooted by something other than the Windows updates but, so far, Sysinternals Rootkit Revealer has shown nothing suspicious.
It's most likely the cruise control already uses a PID algorithm. My guess is there's a reset windup problem, i.e., the integral term is allowed to grow to too high a value.
Just a guess - A PID algorithm is used for cruise control and reset windup isn't handled properly. That is, the integral term (which zeros out the small error that must exist using proportional control alone) can grow large when the system hasn't reached the setpoint after a long time. That condition can occur when driving up a hill or when the car isn't in a suitable gear. A large uncorrected integral term can cause the control system output to saturate (equivalent to flooring the gas pedal) resulting in loss of control. Reset windup is generally handled by logic to halt integration when the controller output reaches a predetermined output.
Roll the printed metal into a cylinder and enclose it in a open-faced box with scrolling wheels at the top and bottom. Mechanically switch to a high gear ratio for fast forward / reverse.
One new feature in Win 7 that seems like a good idea to me is the native ability to do image backups of entire hard drives. I don't think that was available previously without 2nd party apps like Ghost or True Image. The Win 7 native imaging utility is somewhat limited (you can't browse an image for or restore individual files, for example) but it's still useful for disaster recovery.
For experimentation with speed listening, the free (not open source) application from this site can be used: http://www.nch.com.au/scribe/ It's a variable speed constant pitch player aimed at audio transcribers. It can handle mp3, wav, and other formats. The standard version runs under Windows or OSX and the Mini version runs under Windows, OSX, or Linux.
Why not lay the whole book flat in a CAT scanner and grab each page as a vertical slice? I'm kind of joking but maybe someone with CAT scanner experience knows how the machine would respond to the ink on each page and how thin each slice could be.
Slashdot Mirror
I've worked on Mturk since it began in November of 2005 and I've earned $25,000 in all since then. I think the rate of pay is horrible but it's been fun trying to find ways to complete tasks more easily using tools such as GreaseMonkey and custom programming. Out and out robots aren't allowed but tools that reformat the screen for easier reading and clicking and things like that are permitted.
The pay was better when Mturk started but it's been a race to the bottom since then. Still, working from home when I want with almost no responsibility is so appealing that I still do it for the low pay. I can make maybe $5 to $10 per day and spend it on any stupid thing I want without cutting into my regular budget.
http://yourdaily.blogspot.com/2007/03/pc-beer-dispenser.html
Reading TFA, it seems to me that all the safety features are proactive, i.e, they seek to prevent an accident. I doubt that zero-fatalities can be achieved that way. People will ignore the warnings, override the controls, etc. Rather, I think it would be better to concentrate on preventing fatalities when accidents occur using such techniques as enhanced air bags and reliable fire suppression systems. I'd be willing to pay more for those features.
As far as I can tell, the earth shattering news is there's going to be an XML-based TeX successor. Underwhelming, eh?
Disappointing. I feel manipulated, but at least by someone with obvious high intellect.
I stumbled across the Khan lectures a couple of months ago and I think they're great. Many thanks to Salman Khan.
Another series of lectures that appeared on television about 30 years ago is a show called The Mechanical Universe. They're not as deep as some of the Khan lectures but a bit more polished (not that it matters a lot). The Mechanical Universe lectures are available online at:
http://www.learner.org/resources/series42.html
Too bad there isn't more of this kind of material on television. Maybe one of the cable companies could buy the Khan lectures and make them available as an on-demand feature.
Well written. Your post proves your point.
And what do you propose is a productive employment of capital? Reemployment in the casino scheme? Isn't that the most likely thing to happen in practice? Nobody is producing anything.
Your scissors thought experiment is an interesting example of non-physical faster-than-light travel. Along the same lines, here's something I've thought about from time to time and don't know if it makes sense:
Imagine a solid rod that reaches from NY City, NY to San Fransisco, CA. I want to send a signal from NY to CA so I push on the rod and move it one inch toward CA. An observer in CA is watching the rod and when he sees it move he knows I've sent my signal. My question is, how long does it take the signal to reach CA? Doesn't it seem on first thought to be instantaneous? If so, isn't it ftl signaling?
I'm probably overlooking something. Perhaps what is travelling along the bar is a compression wave at some finite speed. But that doesn't seem right because a solid rod, steel for example, doesn't compress an inch when you push on it.
Can anyone tell me what would really happen physically if this experiment was performed?
Thanks.
Not only do flash traders have faster access to data than others, they're given a commission discount by Goldman Sachs because of their volume of trading. If non-insiders tried to make money via arbitrage as the flash traders do, all of their profit would be negated by commissions. The game is definitely rigged.
From the NY Times, May 9, 2010
http://www.nytimes.com/2010/05/08/opinion/08durbin.html
"On Thursday afternoon, the Dow plunged 1,000 points within a few minutes, followed by an equally sudden recovery. We don’t know all the details about the drop, but it was almost certainly the result of computer or human error in a high-speed trading program.
Among the many arcane corners of the financial world highlighted by the Wall Street crisis, high-frequency trading — in which computers scan billions of bits of market data for trading opportunities that may exist for mere fractions of a second — has generated a surprising amount of discussion. Alongside the risk of expensive errors like what happened Thursday, critics say, these programs facilitate insider trading and overwhelm regulators’ access to critical information.
These are fair criticisms. Fortunately, they can also be easily addressed without undermining the positive role that high-frequency trading plays in the market.
Let’s start with the insider trading charge. Often, when an exchange operator receives an investor order and finds that another exchange has a better price, it will “flash” the order to a few select traders in its exchange a split second before sending it to market, giving those traders an opportunity to improve their price, too. When used properly, flashing ensures that investors trade at the best available prices.
But that hair’s breadth of time also gives high-frequency traders an opportunity to make a tidy profit off what amounts to insider information. How? Rather than improve their price, the recipient of a flash can go to the other exchange, buy up all the assets at better prices, and force the original investor to trade with them at an inferior price.
We don’t allow trading based on private knowledge of pending business deals or court rulings, and we shouldn’t allow it in high-frequency trading, either. But that doesn’t mean we should ban flashing all together. Instead, to deter abuse, anyone who gets a preview of a trade, whether by phone or flash, should be required to register with an exchange and keep records of every negotiation."
The review states that VNC isn't firewall friendly but, apparently, the reviewer isn't aware of the single click versions of VNC. These versions run without any changes to the host firewall since the connection is initiated by the host. The single click version of UltraVNC is available at:
http://www.uvnc.com/addons/singleclick.html
I'm wondering what the risks are in using verbal checks (paperless ACH transfers). I pay my monthly electricity bill that way since my power company adds a "convenience charge" for using a credit card. As far as I know, the only thing needed for a verbal check is the account number and bank routing number. What's to stop anyone who knows the account number from issuing a verbal check to themself? The routing number for any bank is available online or by calling the bank. If I dispute a verbal check is the bank required to reverse the charge?
Got a CAT scanner sitting around? Stick the book under it and image each page as a thin slice.
I don't really know if this is technically feasible. How does ink react to xrays? How thin can a image slice be? How fast does a CAT scanner run? Maybe someone with CAT scanner knowledge can fill me in.
I know it's bad form to reply to my own post but I just found the following VirusTotal scan result for a rooted atapi.sys file that causes the blue screen:
http://www.virustotal.com/analisis/85aa49f587f69f30560f02151af2900f3dc71d39d1357727ab41b11ef828a7ff-1265925529
Note that eSafe in this case finds nothing. Doesn't say much for eSafe, huh?
My machines are running ok but I thought I'd look to see if KB977165 which is reported to cause the blue screens was on any of them. It was installed on one machine and not on another. So I thought I'd check a few more things that others say may be causing the problem.
It has been suggested that atapi.sys in the system32/drivers folder might be rootkitted by the update. I compared the SHA-1 hash of atapi.sys on both machines and they were the same (A719156E8AD67456556A02C34E762944234E7A44) so, apparently, update KB977165 didn't change that file in my case.
Some people are saying that atapi.sys is infected with a rootkit. I ran scanned the file through Jotti.com and it found nothing. I also scanned the file at VirusTotal.com and only eSafe reported a problem as follows:
eSafe 7.0.17.0 2010.02.11 Win32.Rootkit
I think this is a false positive since I have identical copies of atapi.sys on both machines and both show the same result on eSafe.
It's possible, I suppose, that I was rooted by something other than the Windows updates but, so far, Sysinternals Rootkit Revealer has shown nothing suspicious.
Do these results agree with anybody elses?
It's most likely the cruise control already uses a PID algorithm. My guess is there's a reset windup problem, i.e., the integral term is allowed to grow to too high a value.
Just a guess -
A PID algorithm is used for cruise control and reset windup isn't handled properly. That is, the integral term (which zeros out the small error that must exist using proportional control alone) can grow large when the system hasn't reached the setpoint after a long time. That condition can occur when driving up a hill or when the car isn't in a suitable gear. A large uncorrected integral term can cause the control system output to saturate (equivalent to flooring the gas pedal) resulting in loss of control. Reset windup is generally handled by logic to halt integration when the controller output reaches a predetermined output.
Roll the printed metal into a cylinder and enclose it in a open-faced box with scrolling wheels at the top and bottom. Mechanically switch to a high gear ratio for fast forward / reverse.
Obviously, when driving behind a smoker, the car exhaust fumes pulled in along with the cigarette smoke don't bother you at all.
Give in to the urge. You're fscked anyway, more or less.
http://www.frozencpu.com/cig-01.html
Here's another Fresnel lens with a mount made for a computer monitor:
http://www.ergoindemand.com/lcd-computer-magnifier-filter.htm
Some gamers claim a Fresnel lens makes the gaming experience more realistic since your eyes remain focused at close to infinity. See, for example:
http://www.rickleephoto.com/rlcoll.htm
One new feature in Win 7 that seems like a good idea to me is the native ability to do image backups of entire hard drives. I don't think that was available previously without 2nd party apps like Ghost or True Image. The Win 7 native imaging utility is somewhat limited (you can't browse an image for or restore individual files, for example) but it's still useful for disaster recovery.
For experimentation with speed listening, the free (not open source) application from this site can be used:
http://www.nch.com.au/scribe/
It's a variable speed constant pitch player aimed at audio transcribers. It can handle mp3, wav, and other formats. The standard version runs under Windows or OSX and the Mini version runs under Windows, OSX, or Linux.
Why not lay the whole book flat in a CAT scanner and grab each page as a vertical slice? I'm kind of joking but maybe someone with CAT scanner experience knows how the machine would respond to the ink on each page and how thin each slice could be.