Well if you read the article they are getting paid north of $50 and hour.
Selective quoting, much? That's the rate for a product launch, which is a short event. Short shifts mean more down-time -- that "$50 an hour" is also "$100 dollars a day" and possibly their only income for a week or a month.
As for the trade shows, the article says Computex models are on $100-$170 for 8 hours, which is £12.50 - £21.25 per hour
That's still not bad as an hourly rate, but again, it's not regular work, and there's a Saturday shift in there too.
But Computex is a big one, and as TFA says, other shows pay $60 a day, bringing us into the same territory as the minimum wage in many US states-- $7.50. That's also less than the UK minimum wage, which is about $9.40.
Let's say you get 3 product launches, the full 5 days at Computex and another 15 days at other trade shows -- that's $2140. Not bad for the equivalent of a month's work, but nowhere near $50ph in real terms. But then, it's not going to be a solid month's work. If that's all you're getting from modelling in a year, that's not even an extra $180 each month over your main source of income.
This is why so many girls go into modelling and find themselves dispirited: it's made to look like a glamorous, well-paid job, but it turns out to be exceptionally sleazy and cheap.
For the most part, girls do not know what they're getting into.
I didn't say commercial product, I said a commercial project. I understood your post completely, and even though it's only to be used internally, as a one-off, it's still for the benefit of a commercial entity. Part of the reason for the low cost of the device is that the Raspberry Pi Foundation aren't putting a commercial markup on it.
Well I have been holding off on a project at work waiting on this thing, and now that its finally "available" I cant get it, I have a red line pressed against me and while I and the bosses are extremely patient, at some point you just got to move on
That time for me was April 24th 2012 at 3PM when I got asked "when are you going to do something about the environmental chamber controls?" 1 order from minibox and a few days later we all scratched it off our list.
PI - 1
So... let's get this straight... A device is released to the public at low margins by a non-profit for the purposes of education, and you're complaining that you couldn't get ahold of one for a commercial project? I think your definition of "people... who have actual real uses for the damn things" is a bit skewed. There were already perfectly appropriate products for your needs on the market.
Yes, but the first run was limited by funds -- they didn't take preorders. It's a pretty generic piece of kit -- most fab plants could knock one up fairly easily. Equipment shortages are usually in either custom components like the Cell processor, or the availability of non-solid-state items like hard drives.
Learning should not be done on a minimalistic system. It simply isn't worth it. Get an old PC for that. You don't want your learning to be constrained by irrelevant factors such as lack of RAM or poor performance or insufficient disk space or unavailable libraries. Get a PC, load a development system and install every development package under the Sun. Your task would be to learn how to code in $foo, not to discover problems with interpreter of $foo on architecture $bar. It is not always easy even for experienced coders to port an already working software from the development system into the embedded target.
I think you're missing the point somewhat. The PC is a heterogenous environment, so you will always have to deal with funny little quirks of compatibility in libraries etc. It's only when you get a homogenous, uniform environment that you stop having to work your way around machine-specific problems.
No, not everything is available on the Raspberry Pi... yet. Yes, someone has to port it. But that's the job of the early adopters, and it only needs done once, and then it is available to everybody.
By the time the first in-a-case Pi comes out, there will no doubt be a hell of a lot of stuff available for it.
The secondary effect will be that there will be better software coverage for all variants of ARM Linux, and Linux users will be able to start migrating away from the Linux i386 and x64 architectures. I've been waiting a long time for desktop Linux to cease to be a PC OS, as it limits its appeal. A desktop ARM Linux would be in direct competition with dumb terminals in the enterprise, and would offer the added bonus of being able to do mixed-mode local and network computing -- maybe they'd still want to use Microsoft Office remotely rather than LibreOffice locally, but they could use Firefox locally without bother.
Not saying its impossible, i have never heard of anyone bricking a phone restoring stock firmware, only '3rd party' ROMs..
That doesn't invalidate the previous poster's point -- you can do what you want with the RPi, including installing any firmware you want, with no risk of bricking the device.
So... sorting is machine learning? MS Word is machine learning? Don't think so.
Nowhere did I nor the GP claim that machines have to be involved. And the machine doesn't use humans in this case, it just uses their choices as its data. So your rebuttal is somewhat unfounded.
Machine learning is learning in the first place, through algorithm: a machine can learn to do a task on its own. Not: a machine assists in a task where someone else learns. In this case, the machine doesn't learn anything. It just acts as a biased dice. The outcome of the process might be called "learned", but the knowledge is in the head of the one that runs the experiment and overlooks the outcome, not in the machine. And the "learning" doesn't generalize, so it doesn't help in improving performance on any other task than selecting between these two designs.
That's why it's not machine learning.
A hell of a lot of machine learning is based around giving the computer equation and let it work out the particular coefficients that give the best possible answer. There are very few machine learning tasks that don't have some sort of experimenter assumptions built in, and no machine learning algorithm is ever 100% generalisable (otherwise machine learning would be a pretty small field, as there would only be one machine learning algorithm!)
The reason that this is classed as a machine learning problem and sort isn't is that a sorting algorithm runs once and gives you a definite answer. But with epsilon-greedy, the computer maintains a theory that approximates the "correct" answer, and over time the answer gets better and better without direct operator control.
Yes, it's a simple algorithm. Yes, you could do a similar thing on paper with a human controller. But that doesn't stop the computer implementation qualifying as machine learning.
In the UK, most places will serve you a medium if you ask for medium rare, simply because most folk who ask for medium rare well send it back to the kitchen because it's "not cooked properly". We're not good with our steaks here.
No, it's not a focus group. A focus group is a bunch of people talking about what they like/don't like. However, humans are very poor at judging what they like. Most living room (en_US "lounge") chairs are uncomfortable. People buy them because when they sit down on them in the showroom, they appear comfortable. Because they encourage poor posture, they take the strain off the sitting muscles. This gives the illusion of relaxation, and tricks people into believing the uncomfortable is comfortable.
A related issue is the fact that the majority of people claim to like their steaks "medium rare". Not because they like them medium rare, but because that's what they hear on the TV.
Focus groups are more often than not a total waste of time.
Indeed, this has no relation to machine learning, whatsoever.
Is there an algorithm? Does the machine use the algorithm to obtain the optimum result? Just because the machine uses humans as its test subjects doesn't stop it being machine learning.
Of course not. The whole point of a focus group is for the facilitator to lead the group to the conclusion he or she wants. Management can't maipulate machine learning algorithms -- only developers can.
And that is precisely why they don't set it to 0/0 = 100%, instead initialising everything to 1:1 = 100%
1(1:1) 2(1:1) 3(1:1)
First user sees 1, clicks it:
1(2:2) 2(1:1) 3(1:1)
At this point, the algorithm could still pick any of the three.
Say it picks 1 again, and this is not clicked:
1(2:3) 2(1:1) 3(1:1)
So say it picks 2 for the next user, but the user doesn't click it:
1(2:3) 2(1:2) 3(1:1)
Well this time it has to pick 3 (unless the 10% random kicks in). Lets assume that's unsuccessful.
1(2:3) 2(1:2) 3(1:2)
OK, so 1 is now favoured, but one more "no click" on 1 levels us off 2:4 = 1:2.
There will never be a true zero probability in the epsilon-greedy algorithm, and it can only approximate zero after accumulating an awful lot of evidence...
The only really bad thing about this approach is that it assumes you don't have a lot of repeat visitors. If you do, they'll be annoyed by seeing different versions, apparently at random (from their perspective).
What he doesn't discuss is what "one" instance of the site is. If you've got tracking cookies switched on, then you can assign one version of the site to the user at first visit and have it persist across browsing sessions.
An oversight on the author's part, but not a huge leap of logic.
Oh FFS -- the use of button colours was what is known in technical jargon as an "example". The point of the article applies to all variables. And while you make think "layout" is less important than "shipping speeds", how do you find out shipping speeds? You have to look for them. If you can't find them, you walk. If you can't find them, chances are it's because of something we call in technical jargon "site design", which includes details such as "layout".
It's easy when you're designing something (I'm guessing you've never had to design anything for the public) to make lots of assumptions without even realising. You might put your "checkout" button where it is on your favourite webshop, but that might actually be the least obvious place to anyone who doesn't already share your shopping habits. Or maybe you think it's a wonderful shade of green, but what you don't realise (as someone with normal site and no understanding of occular defects) is that it's actually invisible against your chosen background to about 5% of the global population.
Epsilon-greedy is one of the most well-known algorithms in machine learning. I'd heard of it before, but I didn't know how it works (I dropped AI after 2nd year), but I do now.
Clever, retort sir, however might I interest you in a long forgotten theory of economics that something bought or sold might possibly have greater value, than that of the mechanism by which it is sold.
Which is why advertising and marketing are such underfunded spheres of public endeavour....
It's funny that scientists try to create ion transistors and DNA-based computers. Nature has found other ways to process information, though. Trying to "replicate" electronic circuitry using biologic systems has all the drawbacks of both approaches and little if any of the benefits.
Science has come up with lots of interesting ideas that have been of no practical value in and of themselves that have turned out to be prerequisites for later innovations. For the most useless of the useless, take Prince Rupert's Drops -- beads of glass that are of no use beyond a simple party trick. And yet there's the possibility of making that if we ever start manufacturing things in space, we might be able to produce perfectly spherical Rupert Drops, practically indestructable ball-bearings.
Your analogy is flawed. If you can't display your photos, you can't use them. It's likely saying it's my fault I got my wallet picked from my pocket because if I put it in a pocket (the best functional place for a wallet) then it's available to anyone who knows how to pick pockets, so I shouldn't expect legal protection.
Then don't leave it somewhere that anyone can pick up and use it.
Put a big friggin watermark over it, anything but posting the image online where it can be copied freely and infinitely at no cost.
Do you think that argument would work if I printed out the Coca-Cola label (available as an image on-line that can be copied freely and infinitely, remember) and stuck it on some cans of supermarket own-brand cola?
Now, this technically is copyright infringement, but when it's for personal use (this case doesn't exactly fit that, but many would argue non-profit use should be included in Fair Use) there is simply no harm to the copyright owner.
Not true. An image loses some of its impact when overused. The image in question is an exceptionally good photograph. If it was part of the brand imaging for one company, it would have a very high impact. If everyone in Houston was making Nya-cat videos with it is a backdrop and every Houston trader used it as their website background, it would lose that impact. With the loss of impact comes loss of value, and the copyright holder can be harmed by loss of earnings if he at any point seeks to monetise the IP assets, as is his right.
He didn't fully automate -- he described it as a "cookie cutter process". Plus the DMCA route meant he had an unambiguous point of contact to deal with -- the ISP. It would have taken a lot of research to find the appropriate people within each company to write a "friendly letter" to.
If he found a 'lot' perhaps he should question whether any real harm is being done. Since he didn't know about the usage yesterday and was seemingly quite happy with everything, finding many instances of illegal use should make you question whether that illegal use has actually caused you any harm.
If you cloned my bank card and took £20 out of my account once a month, every month, from a cash machine I use regularly, I don't think I'd notice the missing money. If I discovered three years in that you'd been doing this, I would expect my 720 quid back. You'd be depriving me of money, just as the copyright violators were depriving the guy in the article of money.
No, not "guilty until proven innocent", not at all. The internet can only operate on the scale it does by eschewing the due diligence that face-to-face business traditionally built itself on. In traditional business, you know a fair bit about clients and suppliers, you've seen their statements and you've got references from other satisfied customers/partners/suppliers.
ISPs know little more than name, address and credit card number for their customers. They have carried out no due diligence, which in the old world would have been considered neglicence. Initial "mere conduit" and "safe harbor" legislation gave the ISPs a free ride -- they had no responsibilities at all. The DMCA improved that by instituting a process whereby the ISPs were given a modicum of responsibility, and forced to seek specific declarations by the alleged infringer.
Perhaps we could call this after-the-fact check-up "overdue diligence"...?
It would have been nice if there had been clauses in there making exceptions for "trusted partners" that encouraged ISPs to do their due diligence ahead of time, but that would have been too easily abused. (I'm thinking big companies using it as effective immunity to DMCA takedown notices.)
Slashdot Mirror
Well if you read the article they are getting paid north of $50 and hour.
Selective quoting, much? That's the rate for a product launch, which is a short event. Short shifts mean more down-time -- that "$50 an hour" is also "$100 dollars a day" and possibly their only income for a week or a month.
As for the trade shows, the article says Computex models are on $100-$170 for 8 hours, which is £12.50 - £21.25 per hour
That's still not bad as an hourly rate, but again, it's not regular work, and there's a Saturday shift in there too.
But Computex is a big one, and as TFA says, other shows pay $60 a day, bringing us into the same territory as the minimum wage in many US states-- $7.50. That's also less than the UK minimum wage, which is about $9.40.
Let's say you get 3 product launches, the full 5 days at Computex and another 15 days at other trade shows -- that's $2140. Not bad for the equivalent of a month's work, but nowhere near $50ph in real terms. But then, it's not going to be a solid month's work. If that's all you're getting from modelling in a year, that's not even an extra $180 each month over your main source of income.
This is why so many girls go into modelling and find themselves dispirited: it's made to look like a glamorous, well-paid job, but it turns out to be exceptionally sleazy and cheap.
For the most part, girls do not know what they're getting into.
I didn't say commercial product, I said a commercial project. I understood your post completely, and even though it's only to be used internally, as a one-off, it's still for the benefit of a commercial entity. Part of the reason for the low cost of the device is that the Raspberry Pi Foundation aren't putting a commercial markup on it.
Well I have been holding off on a project at work waiting on this thing, and now that its finally "available" I cant get it, I have a red line pressed against me and while I and the bosses are extremely patient, at some point you just got to move on
That time for me was April 24th 2012 at 3PM when I got asked "when are you going to do something about the environmental chamber controls?" 1 order from minibox and a few days later we all scratched it off our list.
PI - 1
So... let's get this straight... A device is released to the public at low margins by a non-profit for the purposes of education, and you're complaining that you couldn't get ahold of one for a commercial project? I think your definition of "people ... who have actual real uses for the damn things" is a bit skewed. There were already perfectly appropriate products for your needs on the market.
Yes, but the first run was limited by funds -- they didn't take preorders. It's a pretty generic piece of kit -- most fab plants could knock one up fairly easily. Equipment shortages are usually in either custom components like the Cell processor, or the availability of non-solid-state items like hard drives.
Learning should not be done on a minimalistic system. It simply isn't worth it. Get an old PC for that. You don't want your learning to be constrained by irrelevant factors such as lack of RAM or poor performance or insufficient disk space or unavailable libraries. Get a PC, load a development system and install every development package under the Sun. Your task would be to learn how to code in $foo, not to discover problems with interpreter of $foo on architecture $bar. It is not always easy even for experienced coders to port an already working software from the development system into the embedded target.
I think you're missing the point somewhat. The PC is a heterogenous environment, so you will always have to deal with funny little quirks of compatibility in libraries etc. It's only when you get a homogenous, uniform environment that you stop having to work your way around machine-specific problems.
No, not everything is available on the Raspberry Pi... yet. Yes, someone has to port it. But that's the job of the early adopters, and it only needs done once, and then it is available to everybody.
By the time the first in-a-case Pi comes out, there will no doubt be a hell of a lot of stuff available for it.
The secondary effect will be that there will be better software coverage for all variants of ARM Linux, and Linux users will be able to start migrating away from the Linux i386 and x64 architectures. I've been waiting a long time for desktop Linux to cease to be a PC OS, as it limits its appeal. A desktop ARM Linux would be in direct competition with dumb terminals in the enterprise, and would offer the added bonus of being able to do mixed-mode local and network computing -- maybe they'd still want to use Microsoft Office remotely rather than LibreOffice locally, but they could use Firefox locally without bother.
Not saying its impossible, i have never heard of anyone bricking a phone restoring stock firmware, only '3rd party' ROMs..
That doesn't invalidate the previous poster's point -- you can do what you want with the RPi, including installing any firmware you want, with no risk of bricking the device.
Texas is quite a conservative, family-values state. Not the sort of place that would want to discourage women taking their husbands' names...
The bigger question should be - how do we make life a living hell for this woman?
From her tone, it sounds like it already is. Pity her, don't hate her.
So ... sorting is machine learning? MS Word is machine learning? Don't think so.
Nowhere did I nor the GP claim that machines have to be involved. And the machine doesn't use humans in this case, it just uses their choices as its data. So your rebuttal is somewhat unfounded.
Machine learning is learning in the first place, through algorithm: a machine can learn to do a task on its own. Not: a machine assists in a task where someone else learns. In this case, the machine doesn't learn anything. It just acts as a biased dice. The outcome of the process might be called "learned", but the knowledge is in the head of the one that runs the experiment and overlooks the outcome, not in the machine. And the "learning" doesn't generalize, so it doesn't help in improving performance on any other task than selecting between these two designs.
That's why it's not machine learning.
A hell of a lot of machine learning is based around giving the computer equation and let it work out the particular coefficients that give the best possible answer. There are very few machine learning tasks that don't have some sort of experimenter assumptions built in, and no machine learning algorithm is ever 100% generalisable (otherwise machine learning would be a pretty small field, as there would only be one machine learning algorithm!)
The reason that this is classed as a machine learning problem and sort isn't is that a sorting algorithm runs once and gives you a definite answer. But with epsilon-greedy, the computer maintains a theory that approximates the "correct" answer, and over time the answer gets better and better without direct operator control.
Yes, it's a simple algorithm. Yes, you could do a similar thing on paper with a human controller. But that doesn't stop the computer implementation qualifying as machine learning.
In the UK, most places will serve you a medium if you ask for medium rare, simply because most folk who ask for medium rare well send it back to the kitchen because it's "not cooked properly". We're not good with our steaks here.
No, it's not a focus group. A focus group is a bunch of people talking about what they like/don't like. However, humans are very poor at judging what they like. Most living room (en_US "lounge") chairs are uncomfortable. People buy them because when they sit down on them in the showroom, they appear comfortable. Because they encourage poor posture, they take the strain off the sitting muscles. This gives the illusion of relaxation, and tricks people into believing the uncomfortable is comfortable.
A related issue is the fact that the majority of people claim to like their steaks "medium rare". Not because they like them medium rare, but because that's what they hear on the TV.
Focus groups are more often than not a total waste of time.
Indeed, this has no relation to machine learning, whatsoever.
Is there an algorithm? Does the machine use the algorithm to obtain the optimum result? Just because the machine uses humans as its test subjects doesn't stop it being machine learning.
Of course not. The whole point of a focus group is for the facilitator to lead the group to the conclusion he or she wants. Management can't maipulate machine learning algorithms -- only developers can.
And that is precisely why they don't set it to 0/0 = 100%, instead initialising everything to 1:1 = 100%
1(1:1) 2(1:1) 3(1:1)
First user sees 1, clicks it:
1(2:2) 2(1:1) 3(1:1)
At this point, the algorithm could still pick any of the three.
Say it picks 1 again, and this is not clicked:
1(2:3) 2(1:1) 3(1:1)
So say it picks 2 for the next user, but the user doesn't click it:
1(2:3) 2(1:2) 3(1:1)
Well this time it has to pick 3 (unless the 10% random kicks in). Lets assume that's unsuccessful.
1(2:3) 2(1:2) 3(1:2)
OK, so 1 is now favoured, but one more "no click" on 1 levels us off 2:4 = 1:2.
There will never be a true zero probability in the epsilon-greedy algorithm, and it can only approximate zero after accumulating an awful lot of evidence...
The only really bad thing about this approach is that it assumes you don't have a lot of repeat visitors. If you do, they'll be annoyed by seeing different versions, apparently at random (from their perspective).
What he doesn't discuss is what "one" instance of the site is. If you've got tracking cookies switched on, then you can assign one version of the site to the user at first visit and have it persist across browsing sessions.
An oversight on the author's part, but not a huge leap of logic.
Oh FFS -- the use of button colours was what is known in technical jargon as an "example". The point of the article applies to all variables. And while you make think "layout" is less important than "shipping speeds", how do you find out shipping speeds? You have to look for them. If you can't find them, you walk. If you can't find them, chances are it's because of something we call in technical jargon "site design", which includes details such as "layout".
It's easy when you're designing something (I'm guessing you've never had to design anything for the public) to make lots of assumptions without even realising. You might put your "checkout" button where it is on your favourite webshop, but that might actually be the least obvious place to anyone who doesn't already share your shopping habits. Or maybe you think it's a wonderful shade of green, but what you don't realise (as someone with normal site and no understanding of occular defects) is that it's actually invisible against your chosen background to about 5% of the global population.
Epsilon-greedy is one of the most well-known algorithms in machine learning. I'd heard of it before, but I didn't know how it works (I dropped AI after 2nd year), but I do now.
Clever, retort sir, however might I interest you in a long forgotten theory of economics that something bought or sold might possibly have greater value, than that of the mechanism by which it is sold.
Which is why advertising and marketing are such underfunded spheres of public endeavour....
It's funny that scientists try to create ion transistors and DNA-based computers. Nature has found other ways to process information, though. Trying to "replicate" electronic circuitry using biologic systems has all the drawbacks of both approaches and little if any of the benefits.
Science has come up with lots of interesting ideas that have been of no practical value in and of themselves that have turned out to be prerequisites for later innovations. For the most useless of the useless, take Prince Rupert's Drops -- beads of glass that are of no use beyond a simple party trick. And yet there's the possibility of making that if we ever start manufacturing things in space, we might be able to produce perfectly spherical Rupert Drops, practically indestructable ball-bearings.
Your analogy is flawed. If you can't display your photos, you can't use them. It's likely saying it's my fault I got my wallet picked from my pocket because if I put it in a pocket (the best functional place for a wallet) then it's available to anyone who knows how to pick pockets, so I shouldn't expect legal protection.
Then don't leave it somewhere that anyone can pick up and use it.
Put a big friggin watermark over it, anything but posting the image online where it can be copied freely and infinitely at no cost.
Do you think that argument would work if I printed out the Coca-Cola label (available as an image on-line that can be copied freely and infinitely, remember) and stuck it on some cans of supermarket own-brand cola?
Now, this technically is copyright infringement, but when it's for personal use (this case doesn't exactly fit that, but many would argue non-profit use should be included in Fair Use) there is simply no harm to the copyright owner.
Not true. An image loses some of its impact when overused. The image in question is an exceptionally good photograph. If it was part of the brand imaging for one company, it would have a very high impact. If everyone in Houston was making Nya-cat videos with it is a backdrop and every Houston trader used it as their website background, it would lose that impact. With the loss of impact comes loss of value, and the copyright holder can be harmed by loss of earnings if he at any point seeks to monetise the IP assets, as is his right.
He didn't fully automate -- he described it as a "cookie cutter process". Plus the DMCA route meant he had an unambiguous point of contact to deal with -- the ISP. It would have taken a lot of research to find the appropriate people within each company to write a "friendly letter" to.
If he found a 'lot' perhaps he should question whether any real harm is being done. Since he didn't know about the usage yesterday and was seemingly quite happy with everything, finding many instances of illegal use should make you question whether that illegal use has actually caused you any harm.
If you cloned my bank card and took £20 out of my account once a month, every month, from a cash machine I use regularly, I don't think I'd notice the missing money. If I discovered three years in that you'd been doing this, I would expect my 720 quid back. You'd be depriving me of money, just as the copyright violators were depriving the guy in the article of money.
No, not "guilty until proven innocent", not at all. The internet can only operate on the scale it does by eschewing the due diligence that face-to-face business traditionally built itself on. In traditional business, you know a fair bit about clients and suppliers, you've seen their statements and you've got references from other satisfied customers/partners/suppliers.
ISPs know little more than name, address and credit card number for their customers. They have carried out no due diligence, which in the old world would have been considered neglicence. Initial "mere conduit" and "safe harbor" legislation gave the ISPs a free ride -- they had no responsibilities at all. The DMCA improved that by instituting a process whereby the ISPs were given a modicum of responsibility, and forced to seek specific declarations by the alleged infringer.
Perhaps we could call this after-the-fact check-up "overdue diligence"...?
It would have been nice if there had been clauses in there making exceptions for "trusted partners" that encouraged ISPs to do their due diligence ahead of time, but that would have been too easily abused. (I'm thinking big companies using it as effective immunity to DMCA takedown notices.)