Actually, I found the title quite hilarious. The funniest part is "investigation reveals". You need to be quite an investigator to be able to "reveal" such incredible facts.
This. Maybe, as the article points out, we need some people to work for the government and think about regulation, but I don't see what universities are doing here.
In CS (and AI in particular), a huge part of the research in universities is either obsolete, late compared to the industry, addressing the wrong problems, or just plain wrong. I can't say the same for private research papers ; except a few exceptions, they usually make sense. The reason is simple and is tied to the university research system : in universities, researchers work to publish. Writing papers is the end of their work and it's their job to write as many papers as they can (quantity over quality).
In a company, you write a paper... when you have time, hence only when you have something that is truly ground breaking, the few exception being some companies doing bullshit papers for PR purposes or trying to sell their product, but this is easy to spot, and it is not the case of Apple, Google, Amazon, Facebook,...
I'd be interested to know why they didn't distribute tools to patch the original binaries instead of modified binaries. Maybe copy protection of some sort.
And this is why Equifax need to be sued and go bankrupt, handing all of its money to people who got their information stolen. It won't be much per person, so it is not even fair for the people, but it will be the only fair measure, as well as an important step for the future.
Business school will then teach "Security is important, remember the Equifax case...".
And yes, I'm aware everyone at Equifax will loose their jobs, but that the kind of decision that need to be made.
It's not that hard to believe. In every field, and computer science is one as well, many scientific papers are completely wrong with their analysis, and most of the time, it is pretty hard to detect unless you're an absolute expert in the field. Peer-review is supposed to detect bad papers, but finding experts is not always easy, and sometimes everything you read in the paper seems consistent, so unless you try to reproduce it, you can only say "that looks like good work".
Now, that's not the end of the world. Real discoveries that matter will be replicated and used as a base to develop products that will be effective, whereas crappy stuff will just be ignored.
Yep, had the same reaction. Please show proof. The software is almost publicly available, let's show where in the code there is this "top secret" filter.
Not saying it's true or not, but this is very serious accusation and Kaspersky has explicitely stated that no, they were not that kind of company doing things for the russian secret services. So if they are really lying, please show proofs, everyone wants to know.
But information coming from US Officials is not reliable (or even less) in the Trump era. Any journalist would just relay the information (WSJ or others, they're not scientists, they're clickbaitists) so that doesn't add any value. And there are certainly good reasons for the US government to hurt Kaspersky labs because they are doing security software and they are on the "wrong" side.
But still, the reason why there is no way I would go back to windows after having switched to linux 15 years ago is this : antiviruses. Having a program constantly using your CPU and hard-drive is a nonsense.
And no, I don't buy the "if 99% of people would use linux there would be viruses on Linux too" argument. Many Android phones out there and no one runs an anti-virus (security updates, on the other hand, would be welcome).
The thought of a 3000+ pound hunk of steel driven by an unpredictable person near people on sidewalks is even more insanity. The person could be drunk, texting, having a heart attack, or simply crazy.
Very soon people will realize that having a robot drive is much safer and predictable. Just like they're much better at driving planes, rockets going to space,...
Except Google is not just returning indexed results. It is trying to rank them in order of relevance, and how does it achieve this ? By counting the number of clicks.
And this is a huge problem for Google as well as Youtube and Facebook as well. Because the number of clicks is only related to how well the title seems to answer what the person is looking for, not the actual content, a.k.a. the clickbait effect. There is no way to go back at Google and say "meh, that was crap, forget that I clicked on it".
Youtube introduced likes and dislikes to try to counter that effect but that doesn't work so well.
Now, should you rate the relevance in terms of how much the reader likes the content ? No, that would lead to rate hoaxes higher.
The only solution : give higher relevance to articles that come from verified sources. Not perfect, but better.
This is what I thought when I bought my Lenovo laptop : the laptop is partially paid by all the crapware they install which is fine if you are going to wipe it out.
But the hardware itself revealed being really bad as well. The webcam stopped working after 6 months because the ribbon didn't survive opening/closing the lid, and the plastic overall is crap.
I don't think I'll ever buy another one.
There are other brands like Toshiba that install the same crapware paying part of the laptop, but I've seen similar experience from friends.
It is not as ridiculous as it sounds, and the recent success of Pokemon Go showed it. For games like Mario Kart, there is no reason not to make it available on Android. Yes, not on Xbox as the public is not the right one. There can be some games tied to the console controllers but Mario Kart and such... totally playable on a phone, tablet, or android TV. They're wasting huge amounts of money not doing so.
I've had 0 fraud on my card in Europe in my entire life (20 years of card usage).
10 days after moving to the US, I had a fraud on my debit card with Pin+Chip. And yes, it is much slower than in Europe, but no I don't see the point in extra security since the solution is good enough in Europe (and actually maybe even better).
The real justification for cost is : how much time will phones be updated ?
All the cheap android phones are excellent today, but in two years (at best) you will no longer receive updates, which means basically that you have to buy a new phone if you want to stay secure.
That's the only justification to buy a high-end Galaxy, a Pixel, or even better in this repspect, an iPhone.
The day manufacturers manage to ship stock Android that can get security updates directly from Google, then things will change and I would consider buying a cheap phone, just like I do with cheap laptops running Linux.
Yep. And they used to pollute like hell as well, and kill every person in the car in case of an accident, not like the shit they build today. But yeah, unlike Japanese, electronics made the US cars much less reliable.
The calculator leads more people to believe that they are not overweight!
Sure. Just read again : the BMI does not take fat into account. You can have other tests that are much better (water displacement, pinch testing), but BMI is a rule of thumb that can be wrong for some. The page even mentions it : https://www.nhlbi.nih.gov/heal...
Although BMI can be used for most men and women, it does have some limits:
It may overestimate body fat in athletes and others who have a muscular build.
It may underestimate body fat in older persons and others who have lost muscle.
So before being so heated up, just read what people write.
Another flat out lie, easily disproved if you had bothered to look at.. well.. anything at all instead of bullshitting.
References needed. Since BMI IS a simple rule out of height and weight, they can't do a BMI study other than computing the BMI on the official formula !
But please, if you think otherwise, give some extra links that would show that the BMI statistics are done out of water displacement or pinch testing.
And even your personal experience is consistent with what I'm trying to explain : BMI is a stupid rule of thumb, and for your case, you had to had other tests than the official BMI to prove that you were not overweight.
Let me repeat this. The BMI you link to is a standard index. (Index means... number that has its limited meaning). It is not a "Body fat measurement". It is an index that on average reflects the amount of fat. On average only.
Sure, that is the definition of BMI, immediately followed by a "calculator" taking only the height and weight into account (no fat vs muscle). So his comment is perfectly valid, computing the BMI with this calculator leads to errors and considering overweight people that are not. I doubt that large studies do any differently, since they're likely just doing statistics on weight and height as well.
Now, that is fine since the formula is correct on average (*), which means for an entire population the "muscle" error is compensated, but for a single person, it is -as you mention- an estimate and is therefore wrong.
(*) It still needs to be proven if this formula is correct for all populations. I wouldn't be surprised if a different formula would better apply to e.g. Japan vs Canada.
I would agree, but this comment is basically "I like articles with more than the headline, i.e. text I can read"... while not having read the precise article that had nothing to do about that.
Slashdot Mirror
Even worse : it seems to be a UK thing.
Yellow ... like piss I guess.
Actually, I found the title quite hilarious. The funniest part is "investigation reveals". You need to be quite an investigator to be able to "reveal" such incredible facts.
In the meantime, no one cares about Netflix, Youtube, and others knowing exactly what your are watching and when ...
This. Maybe, as the article points out, we need some people to work for the government and think about regulation, but I don't see what universities are doing here.
In CS (and AI in particular), a huge part of the research in universities is either obsolete, late compared to the industry, addressing the wrong problems, or just plain wrong. I can't say the same for private research papers ; except a few exceptions, they usually make sense. The reason is simple and is tied to the university research system : in universities, researchers work to publish. Writing papers is the end of their work and it's their job to write as many papers as they can (quantity over quality).
In a company, you write a paper ... when you have time, hence only when you have something that is truly ground breaking, the few exception being some companies doing bullshit papers for PR purposes or trying to sell their product, but this is easy to spot, and it is not the case of Apple, Google, Amazon, Facebook, ...
I'd be interested to know why they didn't distribute tools to patch the original binaries instead of modified binaries. Maybe copy protection of some sort.
And this is why Equifax need to be sued and go bankrupt, handing all of its money to people who got their information stolen. It won't be much per person, so it is not even fair for the people, but it will be the only fair measure, as well as an important step for the future.
Business school will then teach "Security is important, remember the Equifax case ...".
And yes, I'm aware everyone at Equifax will loose their jobs, but that the kind of decision that need to be made.
It's not that hard to believe. In every field, and computer science is one as well, many scientific papers are completely wrong with their analysis, and most of the time, it is pretty hard to detect unless you're an absolute expert in the field. Peer-review is supposed to detect bad papers, but finding experts is not always easy, and sometimes everything you read in the paper seems consistent, so unless you try to reproduce it, you can only say "that looks like good work".
Now, that's not the end of the world. Real discoveries that matter will be replicated and used as a base to develop products that will be effective, whereas crappy stuff will just be ignored.
Yeah. Kind of an "auto-fire" button. Twice as fast, guaranteed to wipe out all enemies.
Yep, had the same reaction. Please show proof. The software is almost publicly available, let's show where in the code there is this "top secret" filter.
Not saying it's true or not, but this is very serious accusation and Kaspersky has explicitely stated that no, they were not that kind of company doing things for the russian secret services. So if they are really lying, please show proofs, everyone wants to know.
But information coming from US Officials is not reliable (or even less) in the Trump era. Any journalist would just relay the information (WSJ or others, they're not scientists, they're clickbaitists) so that doesn't add any value. And there are certainly good reasons for the US government to hurt Kaspersky labs because they are doing security software and they are on the "wrong" side.
But still, the reason why there is no way I would go back to windows after having switched to linux 15 years ago is this : antiviruses. Having a program constantly using your CPU and hard-drive is a nonsense.
And no, I don't buy the "if 99% of people would use linux there would be viruses on Linux too" argument. Many Android phones out there and no one runs an anti-virus (security updates, on the other hand, would be welcome).
The thought of a 3000+ pound hunk of steel driven by an unpredictable person near people on sidewalks is even more insanity. The person could be drunk, texting, having a heart attack, or simply crazy.
Very soon people will realize that having a robot drive is much safer and predictable. Just like they're much better at driving planes, rockets going to space, ...
See for yourself. https://github.com/Laurelai/de...
Not really a new path. NVIDIA has the Tegra line for that. It is powering the Nintendo Switch, and other things.
Except Google is not just returning indexed results. It is trying to rank them in order of relevance, and how does it achieve this ? By counting the number of clicks.
And this is a huge problem for Google as well as Youtube and Facebook as well. Because the number of clicks is only related to how well the title seems to answer what the person is looking for, not the actual content, a.k.a. the clickbait effect. There is no way to go back at Google and say "meh, that was crap, forget that I clicked on it".
Youtube introduced likes and dislikes to try to counter that effect but that doesn't work so well.
Now, should you rate the relevance in terms of how much the reader likes the content ? No, that would lead to rate hoaxes higher.
The only solution : give higher relevance to articles that come from verified sources. Not perfect, but better.
This is what I thought when I bought my Lenovo laptop : the laptop is partially paid by all the crapware they install which is fine if you are going to wipe it out.
But the hardware itself revealed being really bad as well. The webcam stopped working after 6 months because the ribbon didn't survive opening/closing the lid, and the plastic overall is crap.
I don't think I'll ever buy another one.
There are other brands like Toshiba that install the same crapware paying part of the laptop, but I've seen similar experience from friends.
So in the end, get a Dell or HP.
It is not as ridiculous as it sounds, and the recent success of Pokemon Go showed it. For games like Mario Kart, there is no reason not to make it available on Android. Yes, not on Xbox as the public is not the right one. There can be some games tied to the console controllers but Mario Kart and such ... totally playable on a phone, tablet, or android TV. They're wasting huge amounts of money not doing so.
Because those suits are only here to address a risk that killed the Soyuz 11 crew : https://en.wikipedia.org/wiki/...
I've had 0 fraud on my card in Europe in my entire life (20 years of card usage).
10 days after moving to the US, I had a fraud on my debit card with Pin+Chip. And yes, it is much slower than in Europe, but no I don't see the point in extra security since the solution is good enough in Europe (and actually maybe even better).
The real justification for cost is : how much time will phones be updated ?
All the cheap android phones are excellent today, but in two years (at best) you will no longer receive updates, which means basically that you have to buy a new phone if you want to stay secure.
That's the only justification to buy a high-end Galaxy, a Pixel, or even better in this repspect, an iPhone.
The day manufacturers manage to ship stock Android that can get security updates directly from Google, then things will change and I would consider buying a cheap phone, just like I do with cheap laptops running Linux.
Yep. And they used to pollute like hell as well, and kill every person in the car in case of an accident, not like the shit they build today. But yeah, unlike Japanese, electronics made the US cars much less reliable.
The calculator leads more people to believe that they are not overweight!
Sure. Just read again : the BMI does not take fat into account. You can have other tests that are much better (water displacement, pinch testing), but BMI is a rule of thumb that can be wrong for some. The page even mentions it : https://www.nhlbi.nih.gov/heal...
Although BMI can be used for most men and women, it does have some limits:
So before being so heated up, just read what people write.
Another flat out lie, easily disproved if you had bothered to look at.. well.. anything at all instead of bullshitting.
References needed. Since BMI IS a simple rule out of height and weight, they can't do a BMI study other than computing the BMI on the official formula !
But please, if you think otherwise, give some extra links that would show that the BMI statistics are done out of water displacement or pinch testing.
And even your personal experience is consistent with what I'm trying to explain : BMI is a stupid rule of thumb, and for your case, you had to had other tests than the official BMI to prove that you were not overweight.
Let me repeat this. The BMI you link to is a standard index. (Index means ... number that has its limited meaning). It is not a "Body fat measurement". It is an index that on average reflects the amount of fat. On average only.
Sure, that is the definition of BMI, immediately followed by a "calculator" taking only the height and weight into account (no fat vs muscle). So his comment is perfectly valid, computing the BMI with this calculator leads to errors and considering overweight people that are not. I doubt that large studies do any differently, since they're likely just doing statistics on weight and height as well.
Now, that is fine since the formula is correct on average (*), which means for an entire population the "muscle" error is compensated, but for a single person, it is -as you mention- an estimate and is therefore wrong.
(*) It still needs to be proven if this formula is correct for all populations. I wouldn't be surprised if a different formula would better apply to e.g. Japan vs Canada.
I would agree, but this comment is basically "I like articles with more than the headline, i.e. text I can read" ... while not having read the precise article that had nothing to do about that.
Nothing is more annoying than people commenting on the title without having read the article.
From the author's UK perspective, I'm not sure if the USA are west or east ... they're in the back.