Shame they couldn't pick up the Netflix/Marvel shows though. Daredevil and Jessica Jones were great, Punisher and Luke Cage were good and even Iron Fist improved a lot in the second season.
What is the politically correct name for stuff like videos encouraging children to kill themselves, or terrorist propaganda videos? If we can't call them toxic what are we supposed to say now?
They stopped doing that in 2017. Aside from anything else there were lawsuits over non-Gmail users having their messages scanned when Gmail users received them. The advertising on Gmail, assuming you don't block it, is now based on data from other Google services you use.
CORRECTION: This post previously identified the sender of the 550 falsely identified URLs as Europolâ(TM)s EU Internet Referral Unit (EU IRU). The sender was in fact, the French national Internet Referral Unit, using Europolâ(TM)s application, which sends the email from an @europol.europa.eu address. The EU IRU has informed us that it is not involved in the national IRUsâ(TM) assessment criteria of terrorist content.
Well it's interesting that Trump has changed his tune from praising Wikileaks, to pretending he doesn't know anything about it. Maybe he knows that Assange does have something on it.
Of course the other problem is that if Assange goes down it further de-ligitimizes Trump's victory on the back of the leaking of Clinton's emails.
There are some good reasons for filtering by gender, some justification for it. Height though seems to be just a thing that conventional beauty considers to be a factor for arbitrary reasons.
The UK government has in the past tried to get assurances that things like solitary confinement punishment won't be used. I don't know if they have always been honoured by the US, certainly some other countries like Egypt have not.
Indefinitely solitary confinement as punishment would be considered torture, for example. Also forced labour is not allowed under UK law, and some states have that I believe (not sure about at the federal level).
It's more a systemic issue, as in if the dating website never recommends any X people, it's re-enforcing that preference. If you see X people on the list it at least presents the opportunity to broaden your options a bit.
It's probably a good thing for the dating site too. There are only so many people who match very specific criteria near you, and making some less perfect match suggestions increases the probability of you finding someone you end up actually liking.
Actually yeah, I take it back. Trans people are in enough danger as it is, best to let them decide to declare trans or not and for people looking to filter trans people if they are that way inclined.
Personally I think gender selection for dating is fine, but with certain limits. It should be the gender the person puts down, nothing like "biological sex at birth" or "has penis". And it should include provision for non-binary.
Assumptions like a specific dopey ass kid is more likely to get their vehicles wrecked than a group of older squares?
That's another interesting example.
In the UK women used to get cheaper car insurance. That turned out to be illegal under gender equality rules. Insurance companies had to stop using gender as a risk factor. Women's car insurance went up, men's went down, in some cases a lot for younger drivers.
I see a lot of complaints about age discrimination in the tech jobs market. I wonder if older tech workers would accept removing age as a risk factor for insurance, pushing their premiums up, in exchange for also removing age discrimination in hiring.
That's the fundamental conflict in play today. Some people think that race is still a factor and we should take action to correct that, in the hope of reaching a post-racial society where it no longer matters. Other think that if we just ignore race completely the problems will fix themselves.
Shouldn't we try to understand why the algorithm came to that conclusion instead of immediately jumping to "check your privilege" ?
Isn't that what we are doing? Trying to understand if there are flaws in the training data or the way the training was administered?
The reality is that it's extremely difficult to provide completely unbiased training data. Worse still it's often an on-going process, e.g. if you have a data set for deciding sentencing it will need to keep evolving over time to reflect new laws and new circumstances that didn't exist before. Just like we need to keep a constant eye on human bias in the system, the same is going to be true of AI.
Which is dangerous if we reduce AI to a black box and can't question it. Hence laws like this.
Do you think that the one hacking charge is all they have? I think it's likely that once they get him they will pile on the charges.
Even if it wasn't Assange, adding extra charges is pretty much SOP for US prosecutors, right? To encourage the accused to take a plea bargain or give up others involved.
Basically it's the level of trust that each gets. CA certs are generally handled transparently without any user interaction and accepted as validating identity. Self signed certs are just used for security and don't prove identity, and some clients may choose to ask the user to confirm their use, typically only once the first time they are encountered.
That gets to the heart of the question. To what extent are preferences something that a person has control over, are something that are dependent on social norms and influences, and which are inherent. And to what extent does re-enforcing those preferences contribute to systemic biases, if any?
Are we okay with filtering people based on those preferences though?
Say you are a significantly below average height male. Conventional standards of attractiveness in many western countries favour tall men. In fact short men are often the butt of jokes. Short men are going to find it hard to meet someone on a dating site if it filters them out based on perceived preferences, or allows the user to set a minimum height when searching.
Some people argue that is just individual preference and it's fine. Sucks but some people are simply less attractive. Incels might agree. But arguably even bothering to list height as an attribute on a person's profile is re-enforcing that standard of conventional attractiveness, which we know influences what people find attractive.
New firefighters have to undergo some tests before being accepted, both physical and academic. Those tests are designed to measure their likely performance at the job.
Just mount a "special rendition" operation once the UK sets him free
Why kidnap someone so high profile from the streets of London when they can just use legal means to do the same thing? Why cause an international incident when all they have to do is put in a normal extradition request?
"Enemy combatant" isn't an actual think, it's just a definition that the US made up in order to avoid giving accused criminals it captured the normal rights accused criminals get.
Much of the information that Wikileaks released is still classified and technically illegal to posses. There is a lot of unreleased material as well, and it's likely that they have been at least trying to compromise his accounts and computers to recover evidence that he has it, which they would argue is classified and can't be used in open court.
Also note that he might not even get a civilian trial. Manning was tried by a military court, as were the inmates at Guantanamo.
If the US has additional charges, and I strongly suspect that they do, they will hold off with them until he is in the US. If they pile them on now it will make it easier for him to argue that he won't get a fair trial or may be tortured (e.g. sent to Guantanamo) over any one of them.
Bias is favoring one thing over another. Which is what you want certain algorithms to do. I want Youtube to find stuff I like. I want Google to find pages that are relevant to me.
How about a dating site that decided you had a bias in favour of tall partners, or light skinned partners? There is no simple answer to this.
Some people consider things like hair colour preference to be a matter of personal taste and completely fine to filter by. Other people complain that they are short and don't get dates and it's unfair.
Out in the real world people may select potential partners based on those preferences. But often relationships start without any deliberate selection, through friends or work or waiting in the same line, and there is less focus on physical attributes. Which is good for people with less conventionally attractive traits, such as being short.
It's something we need to think about, at the very least.
Slashdot Mirror
Shame they couldn't pick up the Netflix/Marvel shows though. Daredevil and Jessica Jones were great, Punisher and Luke Cage were good and even Iron Fist improved a lot in the second season.
What is the politically correct name for stuff like videos encouraging children to kill themselves, or terrorist propaganda videos? If we can't call them toxic what are we supposed to say now?
Disabling comments and up/down votes is suicide for clickbait. It massively down-ranks the video by decreasing the engagement score.
That Verge video only remained visible because so many people were linking to it.
They stopped doing that in 2017. Aside from anything else there were lawsuits over non-Gmail users having their messages scanned when Gmail users received them. The advertising on Gmail, assuming you don't block it, is now based on data from other Google services you use.
From TFA:
CORRECTION: This post previously identified the sender of the 550 falsely identified URLs as Europolâ(TM)s EU Internet Referral Unit (EU IRU). The sender was in fact, the French national Internet Referral Unit, using Europolâ(TM)s application, which sends the email from an @europol.europa.eu address. The EU IRU has informed us that it is not involved in the national IRUsâ(TM) assessment criteria of terrorist content.
So it's actually just the French, not the EU.
Well it's interesting that Trump has changed his tune from praising Wikileaks, to pretending he doesn't know anything about it. Maybe he knows that Assange does have something on it.
Of course the other problem is that if Assange goes down it further de-ligitimizes Trump's victory on the back of the leaking of Clinton's emails.
There are some good reasons for filtering by gender, some justification for it. Height though seems to be just a thing that conventional beauty considers to be a factor for arbitrary reasons.
The UK government has in the past tried to get assurances that things like solitary confinement punishment won't be used. I don't know if they have always been honoured by the US, certainly some other countries like Egypt have not.
"Group X commits more crimes" may be a statistical fact, but should it influence sentencing?
Something being "real" in some sense isn't always the important thing.
Indefinitely solitary confinement as punishment would be considered torture, for example. Also forced labour is not allowed under UK law, and some states have that I believe (not sure about at the federal level).
It's more a systemic issue, as in if the dating website never recommends any X people, it's re-enforcing that preference. If you see X people on the list it at least presents the opportunity to broaden your options a bit.
It's probably a good thing for the dating site too. There are only so many people who match very specific criteria near you, and making some less perfect match suggestions increases the probability of you finding someone you end up actually liking.
Actually yeah, I take it back. Trans people are in enough danger as it is, best to let them decide to declare trans or not and for people looking to filter trans people if they are that way inclined.
Personally I think gender selection for dating is fine, but with certain limits. It should be the gender the person puts down, nothing like "biological sex at birth" or "has penis". And it should include provision for non-binary.
Assumptions like a specific dopey ass kid is more likely to get their vehicles wrecked than a group of older squares?
That's another interesting example.
In the UK women used to get cheaper car insurance. That turned out to be illegal under gender equality rules. Insurance companies had to stop using gender as a risk factor. Women's car insurance went up, men's went down, in some cases a lot for younger drivers.
I see a lot of complaints about age discrimination in the tech jobs market. I wonder if older tech workers would accept removing age as a risk factor for insurance, pushing their premiums up, in exchange for also removing age discrimination in hiring.
That's the fundamental conflict in play today. Some people think that race is still a factor and we should take action to correct that, in the hope of reaching a post-racial society where it no longer matters. Other think that if we just ignore race completely the problems will fix themselves.
Shouldn't we try to understand why the algorithm came to that conclusion instead of immediately jumping to "check your privilege" ?
Isn't that what we are doing? Trying to understand if there are flaws in the training data or the way the training was administered?
The reality is that it's extremely difficult to provide completely unbiased training data. Worse still it's often an on-going process, e.g. if you have a data set for deciding sentencing it will need to keep evolving over time to reflect new laws and new circumstances that didn't exist before. Just like we need to keep a constant eye on human bias in the system, the same is going to be true of AI.
Which is dangerous if we reduce AI to a black box and can't question it. Hence laws like this.
Do you think that the one hacking charge is all they have? I think it's likely that once they get him they will pile on the charges.
Even if it wasn't Assange, adding extra charges is pretty much SOP for US prosecutors, right? To encourage the accused to take a plea bargain or give up others involved.
Basically it's the level of trust that each gets. CA certs are generally handled transparently without any user interaction and accepted as validating identity. Self signed certs are just used for security and don't prove identity, and some clients may choose to ask the user to confirm their use, typically only once the first time they are encountered.
That gets to the heart of the question. To what extent are preferences something that a person has control over, are something that are dependent on social norms and influences, and which are inherent. And to what extent does re-enforcing those preferences contribute to systemic biases, if any?
Are we okay with filtering people based on those preferences though?
Say you are a significantly below average height male. Conventional standards of attractiveness in many western countries favour tall men. In fact short men are often the butt of jokes. Short men are going to find it hard to meet someone on a dating site if it filters them out based on perceived preferences, or allows the user to set a minimum height when searching.
Some people argue that is just individual preference and it's fine. Sucks but some people are simply less attractive. Incels might agree. But arguably even bothering to list height as an attribute on a person's profile is re-enforcing that standard of conventional attractiveness, which we know influences what people find attractive.
New firefighters have to undergo some tests before being accepted, both physical and academic. Those tests are designed to measure their likely performance at the job.
Just mount a "special rendition" operation once the UK sets him free
Why kidnap someone so high profile from the streets of London when they can just use legal means to do the same thing? Why cause an international incident when all they have to do is put in a normal extradition request?
"Enemy combatant" isn't an actual think, it's just a definition that the US made up in order to avoid giving accused criminals it captured the normal rights accused criminals get.
Much of the information that Wikileaks released is still classified and technically illegal to posses. There is a lot of unreleased material as well, and it's likely that they have been at least trying to compromise his accounts and computers to recover evidence that he has it, which they would argue is classified and can't be used in open court.
Also note that he might not even get a civilian trial. Manning was tried by a military court, as were the inmates at Guantanamo.
If the US has additional charges, and I strongly suspect that they do, they will hold off with them until he is in the US. If they pile them on now it will make it easier for him to argue that he won't get a fair trial or may be tortured (e.g. sent to Guantanamo) over any one of them.
Bias is favoring one thing over another. Which is what you want certain algorithms to do. I want Youtube to find stuff I like. I want Google to find pages that are relevant to me.
How about a dating site that decided you had a bias in favour of tall partners, or light skinned partners? There is no simple answer to this.
Some people consider things like hair colour preference to be a matter of personal taste and completely fine to filter by. Other people complain that they are short and don't get dates and it's unfair.
Out in the real world people may select potential partners based on those preferences. But often relationships start without any deliberate selection, through friends or work or waiting in the same line, and there is less focus on physical attributes. Which is good for people with less conventionally attractive traits, such as being short.
It's something we need to think about, at the very least.