Scientific Method
on
Is Sugar Toxic?
·
· Score: 5, Insightful
As a scientist, I like to base my opinions on evidence wherever possible.
Lustig makes strong points which are backed up by studies (cited in his lecture) and are consistent with known biochemical pathways (which he explains).
The vast majority of responses here are complete bunkum: anecdotal evidence, true facts which sound like they are relevant ("you can drown in water!"), and misrepresentation of his central point ("our bodies *need* glucose! He's crazy!")
If you disagree with his position and have evidence to back that up, I'll listen to what you have to say.
Everyone else - you're going to get really frustrated when I don't change my opinion because of what you say.
Let's let evidence and logic have it's moment here.
I used to design software for aircraft pressurization systems, so have come across a lot of information about decompression effects et al.
I was told that wind forces during a fall tend to remove clothing - so that bodies are invariably found in various stages of undress - some completely naked. The victims are unconscious, so take no action to prevent it.
I always thought that was bizarre (which is probably why it has stuck in my memory). Can you confirm or deny this?
1) Stop fighting two wars. Let the people of Iraq settle their own affairs and let Afghanistan rebuild. The threat from terrorism isn't large enough to justify the wars, and there are better ways to deal with that threat.
2) Stop trying to enforce bans on drugs. It's more effective (7x, by some measures) to put the money into education instead of enforcement.
3) Related to #2, let all non-violent drug offenders out of prison immediately. Non violent offenders don't offer a risk to society, and we shouldn't be putting people into prison for what amounts to a personal choice.
4) Close Guantanamo bay. Drop the inmates back in their home country with $50K each and a brand new copy of the Koran. It would be cheaper in the long run than keeping the prison open. Per #1, there are more effective ways to spend money to combat terrorism. (Specifically, money spent to intelligence gathering)
5) Reduce US military presence by 40% worldwide. Give South Korea the task of protecting their Northern border, close bases in Japan and Germany. Keep only a selection of bases worldwide to be used for tactical support, should we ever need to attack a particular area.
6) Get rid of DHS entirely. Route 10% of that budget to the FBI and CIA for intelligence gathering.
7) Stop beating up on illegal immigrants. Make a fast-track for citizenship so that they can start paying taxes as soon as possible. Our population growth is declining. If you don't count immigration, the population growth would be *negative* already. Adding all the immigrants would add almost nothing to overpopulation and add to our economic strength. Oh, and contrary to popular belief, the vast, vast majority of them simply want to be peaceful law-abiding citizens.
The fact that you don't see informed people suggesting things to cut says more about yourself than the situation. Of all the infighting about the national budget, the elephant in the room is our military spending. With the fall of the USSR, and with Cuba and North Korea ludicrous, and with the current changes in the mideast... the world is a much safer place which doesn't need our military presence.
The best preparations are knowledge and experience.
Learn to camp. Join the Boy Scouts or similar when growing up. Learn to fish. Learn to hunt. Go on hikes. Take a first aid course.
Learn to be calm in the face of a completely unfamiliar situation.
You can't really plan for an unexpected event, but you can train yourself to react rationally in unfamiliar circumstances. Having a tendency to improvise a solution will get you much further in an emergency than any preparation for a specific circumstance.
Looks like I typoed the calculations. It's more like 5,000 scans instead of 100,000.
But my point stands. I'd like to see someone who is so certain of the safety show us how little the risk is by taking the challenge (with the correct number of scans).
According to federal sources cited around the web, the amount of radiation is less than 1/1000 times the dosage from 3 minutes of air travel.
If these systems are as safe as you say, being scanned 100,000 times is about the amount of radiation one would get from a 4 hour flight.
It would go a long way towards convincing everyone if, as a publicity stunt, you allowed yourself to be scanned 100,000 times over a four hour period.
The equivalent dosage would be a little less than from a four hour flight, which is a risk that you regularly take as part of your official duties.
If you do this and emerge unharmed, I'm willing to concede the point. Until then, I claim that there is no evidence to warrant any claims as to the safety of airport scanners.
We've all been taught to oppose vigilante actions, and rightly so. We believe that vigilantism is bad at a gut level, and people use that bias to sway public opinion to their own ends.
Vigilantism is when you pass penalty judgement on someone outside of the legal process, for example hanging someone for stealing cattle. The actions of the hackers don't fall under the definition because no one was hurt and no penalties were passed out.
This is simply one group committing a crime in order to expose a much larger crime. We should apply a measure of scale here and realize that the lesser crime can be outweighed by the value to society from exposing the larger issues.
If this is vigilantism, then so is Brad Manning's gift to Wikileaks. Both parties broke the law in order to expose larger crimes which had substantial public interest.
While I don't advocate reporters breaking into places to root around for evidence, at times the public interest is so overwhelming that we can forgive (and even applaud) illegal actions under those limited circumstances.
This is one of those cases. Anonymous can legitimately be labeled with bad words for their actions (immature, hackers, &c), but in the case at hand, their service to society completely outweighs the gravity of their crimes.
The basic problem with any sort of ethics question is that we really don't have a good definition for ethics. Is lying unethical? Is lying unethical if it is for a greater good (such as saving someone's life)? Is depriving someone else opportunity ethical?
The best definition I have seen is based on suffering: if your actions result in more suffering than currently exists, then it's not ethical. If it reduces suffering, then it's ethical.
Thus lying, and many other behaviours don't come under the definition of ethics. Lying is perhaps dishonourable, but it can be used either for good or evil.
People use the fuzzy definition of ethics in order to make you feel guilty and so sacrifice your own well being (and that of your friends) for their agenda (read: "greater good"). In the situation at hand, setting up a foreign call center pulls more people out of suffering than keeping jobs locally.
Those poor people! Anyone who thinks that they are more deserving of a job and comfortable income than poor people are just heartless and cruel! We should outsource *all* our jobs to people who will make better use of the opportunities!
Poppycock.
Your actions can reduce the suffering of 5 people, or 50 people. If you reduce the suffering of 5 people, you are still reducing suffering and so are acting ethically. If choosing the 5 over the 50 helps you out personally, then know that you are still acting ethically despite what some people say.
Do what benefits yourself the most, so long as your actions don't cause suffering. That's all it takes to be ethical - of all the actions to take, eliminate the ones that cause suffering. From what is left, it's OK to receive a benefit.
Rogue admins are extremely rare. So rare that there are many other more likely threats you will encounter, such as hackers or data breach. Worry about those first.
The reality is that most people work in a spirit of cooperation and don't want the black mark on their reputation. They would rather walk away without burning bridges.
That being said, bad admins (and employees in general) spring from two causes: bad treatment and pre-existing jerks.
The best way to handle both situations is to talk to your employees regularly, and find out how they feel. If you know that some policy or other is bothering them, you can avert a crisis very easily if you know about it beforehand.
Some people are just jerks. Don't let these people continue in your organization, even if they are brilliant and highly capable, and even if you don't have an equally brilliant replacement. A mediocre replacement who can work well with others will be much more productive.
(Often said: About 15% of your productivity comes from innate ability, 85% from working with others.)
That having been said, if you're really worried about someone doing you in, make sure you have regular backups and that you personally have access to the backup system. Reformatting a disk and copying data is easy - position yourself so that you can recover completely from the maximum damage they can do.
I can't read *any* of the responses to my post for some reason, but I can see the 1st line and so can get a feel for what's being said.
Intelligence is not well defined, both in common usage and in the fields of psychology and (my field) AI. I agree that there are more aspects which contribute to an overall sense of intelligence.
One of the failings of AI in my mind is the lack of a good definition of intelligence.
As a mathematician, I know what a manifold is, can tell whether something is one, and can construct one to use as an example.
As an AI researcher... nada. There is no consensus in the field as to what intelligence actually *is*. The closest we have is the Turing test, which is not a definition and conflates intelligence with "human intelligence" and "communication".
(So for my own researches I first had to come up with a workable definition, which makes me an outlier in the field.)
Thanks for the PNAS link. I'll read it once more bugs have been squashed in the comment system.
I can't comment on the study because I couldn't find a link to it in the linked article (wtf?).
One of the definitions of intelligence is the ability to put off an immediate reward for a long term benefit. Children are presented with a jelly bean and told "if you can wait until [the researcher] get back, you'll get 3 jelly beans", and then the researcher leaves.
Kids who can put off temptation the longest tend to score highest in IQ tests.
For example, smokers could give up smoking for 3 months and use the money to pay for a high-def TV. This never happens in practice, because of their inability to put off the immediate pleasure in order to get the long-term reward.
BTW, the links on Slashdot have no underlines? With no decoration, you have to mouse around the text in order to see if a link was included in the article.
If the engineer discovers it while typing and corrects it, the cost to fix is pennies (a couple of seconds times the hourly pay of the engineer).
If it's caught on compile, the cost to fix is a buck.
If it's caught on unit test, it's maybe 10 bucks.
If it's caught in code review, it's several people for 10 minutes - maybe 50 bucks.
If it's caught in QA, maybe 100 bucks.
If it's let out into the wild where it inconveniences customers, it can cost many thousands of dollars to fix.
The take-away from this is that efforts made more early in the process are more cost-effective.
Similarly with terrorists. When you mass-screen people you are at the point where one problem can hide among hundreds of thousands of cases, and so the cost of finding the one bad case is very high.
In the previous analysis: 1 in 175,000 Muslims is a terrorist. Consider a tip-line where people can call in to describe suspicious behaviour of Muslims. Will receiving 1 valid tip among 175,000 false alarms be effective?
Rather than pay officers to man the tip line, it's more effective to target your efforts to a small population, where the probabilities of finding a terrorist are much higher.
This is what I mean by useful. A probability which is much more likely to return results, given that we have fixed resources.
The problem with bomb-sniffing dogs has nothing to do with the dogs themselves.
The dogs are supposedly trained to give a "silent" signal to their handler indicating the presence of explosives (or drugs). Something that a citizen wouldn't notice but the trained handler can recognize.
The citizen can't verify whether the dog gave the signal, so it's a classic authentication abuse scenario.
When bomb-sniffing dogs are used, look for lots of false-positives which just happen to be "pretty, large-breasted women" pulled aside for more in-depth screening.
No, it's not a coincidence. It's also not very useful.
You need to examine the distinction between "Statistical Significance" and "Practical Importance".
For example, suppose the difference in IQ scores between people in two cities is 1/2 point. The studies can be extremely accurate and the results can be correct to a strong degree of statistical significance, but the overall result is of no practical importance.
Similar with Muslim extremists. Try to assign a probability (high or low) to each of the following:
Probability that someone is a Muslim, given that they are terrorist. Probability that someone is a terrorist, given that they are Muslim.
There are about 8,500 people on the U.S. "no fly" list, and about 1.5 billion followers of Islam. If *all* terrorists are Muslims, you still have to sort through 175,000 profiled people to find one terrorist.
This is not a piece of data which is useful in and of itself from which to draw conclusions or make policy.
You don't have to be afraid of your neighbors, even if they are are from Pakistan.
This is an example of the "fallacy of the transposed conditional" and how people use it to justify legislation that does nothing to address the problem.
See if you can assign a likelihood (high or low) to the following:
Probability that someone has a laser, given that they shined one at an airplane, Probability that someone shines one at an airplane, given that they have a laser.
The likelihood that anyone having a laser will use it against an airplane is so astronomically small that legislation will have no appreciable effect, but will inconvenience many people.
The logic is precisely backwards, but it sounds like a justification.
Someone should introduce the legislators down under to Bayes Theorem.
Gah! Would it kill you to let us know what the heck it does?
Let's examine the announcement: Dropbox has finally released version 1.0 (but what is it?). The new version comes with hundreds of bug fixes, including invalid file names on Windows, weird Unicode normalizations, Word and Excel file locking, abnormal symlinks hierarchies, and case sensitive file systems on Mac (yeah, but what does it do?). It also adds TrueCrypt support, a Rainbow Shell that offers support for extended attributes, selective sync, a new installation wizard, and reduces resource usage (Awesome! But what does it do?)."
Follow the link and get a great press release. Let's examine *that*:
Huge performance enhancements (but what does it do?) Better user experience (Great! Is it something I could use?) Selective Sync (Also good. Is it useful for something?) Extended Attribute Sync (Another useful feature... or something.)
Follow the link to the Dropbox website, and you find this useful summary:
Our highest quality yet! (Good on you! What's it do?) Huge performance enhancements (Wonderful. Is that important?) Better user experience (Ok, this is just a copy of the press release.)
Go to the Dropbox "about" page, and get all kinds of interesting info:
Dropbox was founded by Drew Houston and Arash Ferdowsi in 2007, and received seed funding from Y Combinator (Academically interesting. What does it do?). Today, Dropbox is well-funded by Sequoia Capital, Accel Partners, and Amidzad (Also good. For what?). Since launching publicly in September of 2008, we've attracted millions of users and are growing rapidly (Growing is good. Do you have a purpose?). We've been featured in the New York Times and on TechCrunch, and have won awards from places like PC Magazine and CNET (Great! Publicity is good. What's it do again?).
Our passion is making a product that rocks and putting it in millions of people's hands (Again, good on you. Still looking for a description of the product...).
If you're interested in joining us, we're looking for more talented people to join the Dropbox team, so be sure to check out our jobs page (Not right now. Can you give me some info on the product?).
Going to the home page is equally enlightening. An enormous button invites me download the product, or watch a video of some sort.
Sorry guys, but I don't download something unless I at least know what the heck it's supposed to do.
I was trying to figure out whether any of the safety claims (for or against) are true.
Guess what? You can't. The press, blogosphere, and government has made such a pigs dinner of the situation that it's impossible to make heads or tails of the safety claims.
Nothing is compelling either way. We could just as well use a Ouija board to assess the safety.
In March of this year (2010), I did an image search of the most common English words and visually counted into the list to the first porn image. (For my subjective definition of porn).
My reasoning was that some words are closely associated with prurient intent, and I was wondering if there could be a "distance to porn" defined in conceptual space. For example, "foot" would probably have a large prurient distance, while "head" would be a short distance due to phrases like "giving head" &c.
To my surprise, nearly all common English words are really, *really* close to porn by that measure.
It's 8 months later. I redo the search and see almost nothing which would constitute porn.
I guess they've cleaned up their image search algorithm or something.
Here's the top of original list, from March 2010:
Rank::Word::Dist(Porn) 1 the 88 2 of 47 3 to 65 4 and 24 5 a 35 6 in 22 7 is 48 8 it 53 9 you 40 10 that 28 11 he 39 12 was 21 13 for 64 14 on 21 15 are 78 16 with 27 17 as 23 18 I 22 19 his 21 20 they 31 21 be 53 22 at 21 23 one 32 24 have 57 25 this 33 26 from 36 27 or 24 28 had 62 29 by 21 30 hot 23 31 but 24 32 some 33 33 what 45 34 there 29 35 we 24 36 can 82 37 out 44 38 other 87 39 were 43 40 all 23
This is a common misconception about holograms which has come about because of movie special effects.
A real hologram can show the illusion of something floating in front of you, but only so long as your gaze is directed *at* the hologram. Thus, a hologram "picture" hanging on the wall can only show an object while you're looking at the picture, but direct your gaze to the wall left or right and you see the wall. You can see a little bit around the object, but you can't walk around and see behind it because then you would be looking away from the hologram.
For a complete 3-d image you need a "band" of hologram that goes all around the room. Now, wherever you look you are looking into the hologram, and will see the image at the corresponding angle. The requirement to be looking at the hologram is still there - you can't look down through the object to the floor.
If the hologram covered every surface of the room you could have a the illusion of a complete 3-d representation of an object. In this case you could walk around it and view it from any angle, including from below and from above.
However, if another person were in the room with you, you could not see the object if they were between you and the wall. If they are opposite the image from you then you will see them, not the object. If you and they are at 90 degrees to the object, then you can see the object... but you can only walk around it to the point where the other person obscures your view of the wall.
Holograms don't cause light to change direction in mid air. It's just an optical effect that 'kinda reverses the focus in a way that tricks your eyes into thinking there's an object there.
Slashdot Mirror
As a scientist, I like to base my opinions on evidence wherever possible.
Lustig makes strong points which are backed up by studies (cited in his lecture) and are consistent with known biochemical pathways (which he explains).
The vast majority of responses here are complete bunkum: anecdotal evidence, true facts which sound like they are relevant ("you can drown in water!"), and misrepresentation of his central point ("our bodies *need* glucose! He's crazy!")
If you disagree with his position and have evidence to back that up, I'll listen to what you have to say.
Everyone else - you're going to get really frustrated when I don't change my opinion because of what you say.
Let's let evidence and logic have it's moment here.
I used to design software for aircraft pressurization systems, so have come across a lot of information about decompression effects et al.
I was told that wind forces during a fall tend to remove clothing - so that bodies are invariably found in various stages of undress - some completely naked. The victims are unconscious, so take no action to prevent it.
I always thought that was bizarre (which is probably why it has stuck in my memory). Can you confirm or deny this?
Brian Tracy has a series of lectures about exactly that point (selling).
If you list the good features of your product no one will care. Worse, no one will give you the 10 minutes you need to describe the product.
Example1:
a) Our copier can produce 40 pages per minute ...has a 500 GB disk ...has networking capabilities ...can sort, collate, and staple
b)
c)
d)
If you put the product in terms which are advantageous to the listener (usually money), then you spark their interest.
Example2:
Mr manager, if you purchase our copier you can reduce your costs by $2000 per month.
Would you like me to explain how?
It's obvious when you think about it.
Okay, I'll bite.
1) Stop fighting two wars. Let the people of Iraq settle their own affairs and let Afghanistan rebuild. The threat from terrorism isn't large enough to justify the wars, and there are better ways to deal with that threat.
2) Stop trying to enforce bans on drugs. It's more effective (7x, by some measures) to put the money into education instead of enforcement.
3) Related to #2, let all non-violent drug offenders out of prison immediately. Non violent offenders don't offer a risk to society, and we shouldn't be putting people into prison for what amounts to a personal choice.
4) Close Guantanamo bay. Drop the inmates back in their home country with $50K each and a brand new copy of the Koran. It would be cheaper in the long run than keeping the prison open. Per #1, there are more effective ways to spend money to combat terrorism. (Specifically, money spent to intelligence gathering)
5) Reduce US military presence by 40% worldwide. Give South Korea the task of protecting their Northern border, close bases in Japan and Germany. Keep only a selection of bases worldwide to be used for tactical support, should we ever need to attack a particular area.
6) Get rid of DHS entirely. Route 10% of that budget to the FBI and CIA for intelligence gathering.
7) Stop beating up on illegal immigrants. Make a fast-track for citizenship so that they can start paying taxes as soon as possible. Our population growth is declining. If you don't count immigration, the population growth would be *negative* already. Adding all the immigrants would add almost nothing to overpopulation and add to our economic strength. Oh, and contrary to popular belief, the vast, vast majority of them simply want to be peaceful law-abiding citizens.
The fact that you don't see informed people suggesting things to cut says more about yourself than the situation. Of all the infighting about the national budget, the elephant in the room is our military spending. With the fall of the USSR, and with Cuba and North Korea ludicrous, and with the current changes in the mideast... the world is a much safer place which doesn't need our military presence.
The best preparations are knowledge and experience.
Learn to camp. Join the Boy Scouts or similar when growing up. Learn to fish. Learn to hunt. Go on hikes. Take a first aid course.
Learn to be calm in the face of a completely unfamiliar situation.
You can't really plan for an unexpected event, but you can train yourself to react rationally in unfamiliar circumstances. Having a tendency to improvise a solution will get you much further in an emergency than any preparation for a specific circumstance.
Post in haste, regret at leisure.
Looks like I typoed the calculations. It's more like 5,000 scans instead of 100,000.
But my point stands. I'd like to see someone who is so certain of the safety show us how little the risk is by taking the challenge (with the correct number of scans).
Mr. Pistole:
According to federal sources cited around the web, the amount of radiation is less than 1/1000 times the dosage from 3 minutes of air travel.
If these systems are as safe as you say, being scanned 100,000 times is about the amount of radiation one would get from a 4 hour flight.
It would go a long way towards convincing everyone if, as a publicity stunt, you allowed yourself to be scanned 100,000 times over a four hour period.
The equivalent dosage would be a little less than from a four hour flight, which is a risk that you regularly take as part of your official duties.
If you do this and emerge unharmed, I'm willing to concede the point. Until then, I claim that there is no evidence to warrant any claims as to the safety of airport scanners.
Rajstennaj Barrabas
Nashua, NH
The problem is one of definition and scale.
We've all been taught to oppose vigilante actions, and rightly so. We believe that vigilantism is bad at a gut level, and people use that bias to sway public opinion to their own ends.
Vigilantism is when you pass penalty judgement on someone outside of the legal process, for example hanging someone for stealing cattle. The actions of the hackers don't fall under the definition because no one was hurt and no penalties were passed out.
This is simply one group committing a crime in order to expose a much larger crime. We should apply a measure of scale here and realize that the lesser crime can be outweighed by the value to society from exposing the larger issues.
If this is vigilantism, then so is Brad Manning's gift to Wikileaks. Both parties broke the law in order to expose larger crimes which had substantial public interest.
While I don't advocate reporters breaking into places to root around for evidence, at times the public interest is so overwhelming that we can forgive (and even applaud) illegal actions under those limited circumstances.
This is one of those cases. Anonymous can legitimately be labeled with bad words for their actions (immature, hackers, &c), but in the case at hand, their service to society completely outweighs the gravity of their crimes.
The basic problem with any sort of ethics question is that we really don't have a good definition for ethics. Is lying unethical? Is lying unethical if it is for a greater good (such as saving someone's life)? Is depriving someone else opportunity ethical?
The best definition I have seen is based on suffering: if your actions result in more suffering than currently exists, then it's not ethical. If it reduces suffering, then it's ethical.
Thus lying, and many other behaviours don't come under the definition of ethics. Lying is perhaps dishonourable, but it can be used either for good or evil.
People use the fuzzy definition of ethics in order to make you feel guilty and so sacrifice your own well being (and that of your friends) for their agenda (read: "greater good"). In the situation at hand, setting up a foreign call center pulls more people out of suffering than keeping jobs locally.
Those poor people! Anyone who thinks that they are more deserving of a job and comfortable income than poor people are just heartless and cruel! We should outsource *all* our jobs to people who will make better use of the opportunities!
Poppycock.
Your actions can reduce the suffering of 5 people, or 50 people. If you reduce the suffering of 5 people, you are still reducing suffering and so are acting ethically. If choosing the 5 over the 50 helps you out personally, then know that you are still acting ethically despite what some people say.
Do what benefits yourself the most, so long as your actions don't cause suffering. That's all it takes to be ethical - of all the actions to take, eliminate the ones that cause suffering. From what is left, it's OK to receive a benefit.
Rogue admins are extremely rare. So rare that there are many other more likely threats you will encounter, such as hackers or data breach. Worry about those first.
The reality is that most people work in a spirit of cooperation and don't want the black mark on their reputation. They would rather walk away without burning bridges.
That being said, bad admins (and employees in general) spring from two causes: bad treatment and pre-existing jerks.
The best way to handle both situations is to talk to your employees regularly, and find out how they feel. If you know that some policy or other is bothering them, you can avert a crisis very easily if you know about it beforehand.
Some people are just jerks. Don't let these people continue in your organization, even if they are brilliant and highly capable, and even if you don't have an equally brilliant replacement. A mediocre replacement who can work well with others will be much more productive.
(Often said: About 15% of your productivity comes from innate ability, 85% from working with others.)
That having been said, if you're really worried about someone doing you in, make sure you have regular backups and that you personally have access to the backup system. Reformatting a disk and copying data is easy - position yourself so that you can recover completely from the maximum damage they can do.
Does this mean we can get their IPV4 addresses back?
Just 'sayin
I can't read *any* of the responses to my post for some reason, but I can see the 1st line and so can get a feel for what's being said.
Intelligence is not well defined, both in common usage and in the fields of psychology and (my field) AI. I agree that there are more aspects which contribute to an overall sense of intelligence.
One of the failings of AI in my mind is the lack of a good definition of intelligence.
As a mathematician, I know what a manifold is, can tell whether something is one, and can construct one to use as an example.
As an AI researcher... nada. There is no consensus in the field as to what intelligence actually *is*. The closest we have is the Turing test, which is not a definition and conflates intelligence with "human intelligence" and "communication".
(So for my own researches I first had to come up with a workable definition, which makes me an outlier in the field.)
Thanks for the PNAS link. I'll read it once more bugs have been squashed in the comment system.
I can't comment on the study because I couldn't find a link to it in the linked article (wtf?).
One of the definitions of intelligence is the ability to put off an immediate reward for a long term benefit. Children are presented with a jelly bean and told "if you can wait until [the researcher] get back, you'll get 3 jelly beans", and then the researcher leaves.
Kids who can put off temptation the longest tend to score highest in IQ tests.
For example, smokers could give up smoking for 3 months and use the money to pay for a high-def TV. This never happens in practice, because of their inability to put off the immediate pleasure in order to get the long-term reward.
BTW, the links on Slashdot have no underlines? With no decoration, you have to mouse around the text in order to see if a link was included in the article.
Love
The
New
Site.
Whitespace
good!!
Keep
Up
The
Good
Work!
Suppose there is a bug in some code.
If the engineer discovers it while typing and corrects it, the cost to fix is pennies (a couple of seconds times the hourly pay of the engineer).
If it's caught on compile, the cost to fix is a buck.
If it's caught on unit test, it's maybe 10 bucks.
If it's caught in code review, it's several people for 10 minutes - maybe 50 bucks.
If it's caught in QA, maybe 100 bucks.
If it's let out into the wild where it inconveniences customers, it can cost many thousands of dollars to fix.
The take-away from this is that efforts made more early in the process are more cost-effective.
Similarly with terrorists. When you mass-screen people you are at the point where one problem can hide among hundreds of thousands of cases, and so the cost of finding the one bad case is very high.
In the previous analysis: 1 in 175,000 Muslims is a terrorist. Consider a tip-line where people can call in to describe suspicious behaviour of Muslims. Will receiving 1 valid tip among 175,000 false alarms be effective?
Rather than pay officers to man the tip line, it's more effective to target your efforts to a small population, where the probabilities of finding a terrorist are much higher.
This is what I mean by useful. A probability which is much more likely to return results, given that we have fixed resources.
The problem with bomb-sniffing dogs has nothing to do with the dogs themselves.
The dogs are supposedly trained to give a "silent" signal to their handler indicating the presence of explosives (or drugs). Something that a citizen wouldn't notice but the trained handler can recognize.
The citizen can't verify whether the dog gave the signal, so it's a classic authentication abuse scenario.
When bomb-sniffing dogs are used, look for lots of false-positives which just happen to be "pretty, large-breasted women" pulled aside for more in-depth screening.
No, it's not a coincidence. It's also not very useful.
You need to examine the distinction between "Statistical Significance" and "Practical Importance".
For example, suppose the difference in IQ scores between people in two cities is 1/2 point. The studies can be extremely accurate and the results can be correct to a strong degree of statistical significance, but the overall result is of no practical importance.
Similar with Muslim extremists. Try to assign a probability (high or low) to each of the following:
Probability that someone is a Muslim, given that they are terrorist.
Probability that someone is a terrorist, given that they are Muslim.
There are about 8,500 people on the U.S. "no fly" list, and about 1.5 billion followers of Islam. If *all* terrorists are Muslims, you still have to sort through 175,000 profiled people to find one terrorist.
This is not a piece of data which is useful in and of itself from which to draw conclusions or make policy.
You don't have to be afraid of your neighbors, even if they are are from Pakistan.
This is an example of the "fallacy of the transposed conditional" and how people use it to justify legislation that does nothing to address the problem.
See if you can assign a likelihood (high or low) to the following:
Probability that someone has a laser, given that they shined one at an airplane,
Probability that someone shines one at an airplane, given that they have a laser.
The likelihood that anyone having a laser will use it against an airplane is so astronomically small that legislation will have no appreciable effect, but will inconvenience many people.
The logic is precisely backwards, but it sounds like a justification.
Someone should introduce the legislators down under to Bayes Theorem.
Yours is the clearest explanation I've found so far, so I've copy/pasted it to my site (with attribution).
Thanks.
The notation doesn't translate well in text, so I'm supplying a link.
Can you explain this one?
Another puzzle.
Enjoy!
Gah! Would it kill you to let us know what the heck it does?
Let's examine the announcement: Dropbox has finally released version 1.0 (but what is it?). The new version comes with hundreds of bug fixes, including invalid file names on Windows, weird Unicode normalizations, Word and Excel file locking, abnormal symlinks hierarchies, and case sensitive file systems on Mac (yeah, but what does it do?). It also adds TrueCrypt support, a Rainbow Shell that offers support for extended attributes, selective sync, a new installation wizard, and reduces resource usage (Awesome! But what does it do?)."
Follow the link and get a great press release. Let's examine *that*:
Huge performance enhancements (but what does it do?) Better user experience (Great! Is it something I could use?) Selective Sync (Also good. Is it useful for something?) Extended Attribute Sync (Another useful feature... or something.)
Follow the link to the Dropbox website, and you find this useful summary:
Our highest quality yet! (Good on you! What's it do?) Huge performance enhancements (Wonderful. Is that important?) Better user experience (Ok, this is just a copy of the press release.)
Go to the Dropbox "about" page, and get all kinds of interesting info:
Dropbox was founded by Drew Houston and Arash Ferdowsi in 2007, and received seed funding from Y Combinator (Academically interesting. What does it do?). Today, Dropbox is well-funded by Sequoia Capital, Accel Partners, and Amidzad (Also good. For what?). Since launching publicly in September of 2008, we've attracted millions of users and are growing rapidly (Growing is good. Do you have a purpose?). We've been featured in the New York Times and on TechCrunch, and have won awards from places like PC Magazine and CNET (Great! Publicity is good. What's it do again?).
Our passion is making a product that rocks and putting it in millions of people's hands (Again, good on you. Still looking for a description of the product...).
If you're interested in joining us, we're looking for more talented people to join the Dropbox team, so be sure to check out our jobs page (Not right now. Can you give me some info on the product?).
Going to the home page is equally enlightening. An enormous button invites me download the product, or watch a video of some sort.
Sorry guys, but I don't download something unless I at least know what the heck it's supposed to do.
I was trying to figure out whether any of the safety claims (for or against) are true.
Guess what? You can't. The press, blogosphere, and government has made such a pigs dinner of the situation that it's impossible to make heads or tails of the safety claims.
Nothing is compelling either way. We could just as well use a Ouija board to assess the safety.
Here's my analysis.
If you agree or disagree [about safety claims] and have insight, I'd like to hear it.
Jeri Ellsworth, AKA "Lady Ada"
Read some of her articles on hackaday.
Brilliant, clever, and resourceful. Definitely hero material.
In March of this year (2010), I did an image search of the most common English words and visually counted into the list to the first porn image. (For my subjective definition of porn).
My reasoning was that some words are closely associated with prurient intent, and I was wondering if there could be a "distance to porn" defined in conceptual space. For example, "foot" would probably have a large prurient distance, while "head" would be a short distance due to phrases like "giving head" &c.
To my surprise, nearly all common English words are really, *really* close to porn by that measure.
It's 8 months later. I redo the search and see almost nothing which would constitute porn.
I guess they've cleaned up their image search algorithm or something.
Here's the top of original list, from March 2010:
Rank::Word::Dist(Porn)
1 the 88
2 of 47
3 to 65
4 and 24
5 a 35
6 in 22
7 is 48
8 it 53
9 you 40
10 that 28
11 he 39
12 was 21
13 for 64
14 on 21
15 are 78
16 with 27
17 as 23
18 I 22
19 his 21
20 they 31
21 be 53
22 at 21
23 one 32
24 have 57
25 this 33
26 from 36
27 or 24
28 had 62
29 by 21
30 hot 23
31 but 24
32 some 33
33 what 45
34 there 29
35 we 24
36 can 82
37 out 44
38 other 87
39 were 43
40 all 23
This is a common misconception about holograms which has come about because of movie special effects.
A real hologram can show the illusion of something floating in front of you, but only so long as your gaze is directed *at* the hologram. Thus, a hologram "picture" hanging on the wall can only show an object while you're looking at the picture, but direct your gaze to the wall left or right and you see the wall. You can see a little bit around the object, but you can't walk around and see behind it because then you would be looking away from the hologram.
For a complete 3-d image you need a "band" of hologram that goes all around the room. Now, wherever you look you are looking into the hologram, and will see the image at the corresponding angle. The requirement to be looking at the hologram is still there - you can't look down through the object to the floor.
If the hologram covered every surface of the room you could have a the illusion of a complete 3-d representation of an object. In this case you could walk around it and view it from any angle, including from below and from above.
However, if another person were in the room with you, you could not see the object if they were between you and the wall. If they are opposite the image from you then you will see them, not the object. If you and they are at 90 degrees to the object, then you can see the object... but you can only walk around it to the point where the other person obscures your view of the wall.
Holograms don't cause light to change direction in mid air. It's just an optical effect that 'kinda reverses the focus in a way that tricks your eyes into thinking there's an object there.