The thing is, there's forced and then there's forced. You can be forced to do something even if nobody's holding a gun to your head. See "Hobson's choice": you can choose the first horse in line, or you can go without a horse. You aren't forced to choose that particular horse, but if you want a horse you don't have any other choices. You could go to another stable, assuming that a) there's another stable in town and b) it still has any horses left. If I have the biggest stable in town and any other stables are stuck with having only a couple of horses because I'm buying up all but a fraction of the feed, you'll likely find yourself without any other options. And is it really a choice if your only other choice is "do without"?
That's where Facebook is. When they weren't a monopoly and were facing competition, they went and used their money to buy up enough competitors that a critical mass of users wound up on their platform like it or not. Once that critical mass was reached, anyone looking at choices in social-media platforms was faced with the realization that they really didn't have a choice. The majority of their friends were on Facebook, so their choices were to use Facebook or to not interact with their friends on social media. Hobson's choice.
My guess would be that the "collateral damage" is related to IPv4 address-space exhaustion. VPN providers discard "dirty" blocks of address space (ones flagged as being associated with VPNs and thus blacklisted) and move to new "clean" blocks. Used to be that those discarded blocks would take quite a while to be reallocated to a new owner, but now with IPv4 space being scarce they're being reallocated to new owners soon after being made available. The new owners now find themselves with address space that's on various blacklists because of the previous owners, but with no idea about the situation because they don't know who the previous owner even was let alone who's got those netblocks on blacklists.
The only fix there would lie with Netflix, not the ISPs or whoever got the "dirty" address space, because it's Netflix's blacklist and they're the only ones who can update it. They'd have to monitor address space releases and reallocations and remove stale entries from their blacklist, and I doubt they want to make that effort as long as they can get away with pointing fingers at someone else.
That'd be my reaction too: "Why would my ISP know anything about why you guys added an address block to your blacklist, or even when you added it to know if they even owned that block at the time?".
I actually think Biden's right, the trigger that sets things in motion for the next major war will be some sort of cyber-attack. Not because the US or any other power intends to start a shooting war over it, but just because that event started a chain of misunderstandings and judgemental errors that will result in the shooting starting. Remember that the last time we were on the verge of a nuclear war (1995, Boris Yeltsin had actually retrieved the launch codes and was prepared to issue launch orders) it was because of the launch of a Norwegian research rocket. And Russia had even been told about the launch beforehand, the information simply hadn't been passed up the chain of command. My guess is it'll be the same thing: a cyber-attack (or more than one) gets everyone finger-pointing and causes disruptions that could be taken advantage of, then something else happens on one side that gets misinterpreted by the other side, someone in the chain decides that if they wait long enough for solid confirmation it'll be too late to act, or a couple of opposing pilots get a little too rambunctious and an accidental missile release triggers itchy trigger fingers on both sides, or an innocent vehicle gets misidentified as a threat and shot which triggers retaliation...
Another option would be to simply exercise Congress's authority under USC Article 3 section 1 to establish courts other than the Supreme Court. Make APJ judges actual judges, their proceedings actual court proceedings, and just to mess with the Justices make those courts directly beneath the Supreme Court itself so any appeals of their judgements go straight to the Justices for consideration. I don't even see anything in the Constitution that requires Congress to require confirmation of judges when establishing the courts, since they wouldn't be officers of the executive branch but be part of the judicial branch even if appointed by the director of an executive agency.
It seems like the Supremes are determined to eliminate any non-political or insulated-from-politics positions in the executive branch and return it completely to the spoils system.
The question the US government needs to ask themselves is this: can the government of Iran seize the domains of US news outlets to prevent them from spreading what the Iranian government considers disinformation? I mean, the Iranian authorities can get the same kind of court order the US government got and all, so why would their seizure be any less legitimate?
One wonders what's going to happen one of these days in a state with a "stand your ground" law when someone notices armed intruders (the SWAT team) on a doorcam and proceeds to deal with said intruders as the law allows for. Sure they have body armor, but .30-caliber rounds from a hunting rifle will penetrate that easily and single-slug deer-hunting rounds from a shotgun won't even need to penetrate.
Personally I think things like this are a good thing. Making the whole copyright mess cause actual financial damage to large corporations (by eg. having a triple-A mainstream game title flop because gamers can't safely promote it to others) may be the only way to get the mess fixed.
And of course the question of whether the web site in question even had any authentication set up, or if it (like so many) simply assumed nobody would ever use anything other than their real name to submit a quote. As one judge has pointed out, "not desired" doesn't equate to "unauthorized".
And this is going to remain the case as long as platforms are liable for failing to remove content but not liable for wrongfully removing content, and copyright holders are entitled to damages for proven infringement but not liable for damages for wrongful claims of infringement.
One addition I'd love to make to the DMCA is that if a counter-notice is filed and the copyright holder fails to file an infringement suit within 90 days of being informed of the counter-notice, they are barred from ever bringing any further claims for the items involved (including future DMCA notices) and are automatically liable for statutory damages equal to the damages that would have resulted from their claimed infringement.
By that logic the camera on the cel phone used should render the phone manufacturer liable. After all, it was the phone manufacturer, not the user, that put the camera in there and made the phone. And surely they knew that, since that camera can be used to document how fast you're travelling, teenagers would use it to document themselves going for exactly those same high speed numbers to show off just as you say. Similarly for the car manufacturers, knowing that simply putting a speedometer in the car would encourage reckless teenagers to capture an image of it to demonstrate just how fast they were capable of going.
I prefer the methods outlined in "David's Sling" in the data duels. Rather than removing the bad information, it's tagged with details of it's problems, links to supporting sources for correct information and the results of applying the corrected information to the argument the bad information was supposed to be in support of. This allows a response to the "think/do-the-research for yourself" line from the conspiracy theorists: "We did, and here's what we came up with.".
I really think the courts need to rethink their standards for sanctions. It's one thing when the defendant has to bring up things not mentioned in the complaint to get the complaint dismissed. But when the defense can win a motion for dismissal based solely on what's in the complaint, not sanctioning the plaintiff's attorney should be the exceptional case.
At the motion-to-dismiss stage the judge can rule only on matters of law requiring no determination of facts, and any ambiguity has to be taken in the light most favorable to the plaintiff. Given that, if the judge can find from the plaintiff's own complaint that the complaint doesn't state a case then the plaintiff's attorney certainly should have been able to come to the same conclusion. So either they filed it knowing it was baseless (which they're required not to do) or they filed it without doing their research on what the law and precedent say about it (which is also an ethical no-no) or they need to go back to law school. If it's really a valid complaint then they should be able to find at least one case that's still good law with a set of facts close enough to theirs to require a determination of facts to decide, which would let the complaint survive the motion to dismiss.
The biggest problem with privacy issues is that the Internet companies tend to give you Hobson's choice: you either accept them doing whatever they want or you don't use their service at all, even when it's entirely possible for the service to work fine without collecting all that user data. Mostly that happens because they've settled on a business model that depends on selling that data for their revenue, without that they have no other source of revenue.
I don't like the idea of legislating business models, it tends to be too easy to abuse that power. For these cases I don't think it's necessary though. Just legislate liability: any company that collects personal information is 100% liable for the actual costs or a statutory damage amount, whichever is greater, plus legal costs and fees for every person whose data is compromised where that compromise can be traced back to their collection of the data. That means that if company A collects data and sells it to a data broker and that data broker's systems get compromised the data broker would be liable <i>and</i> company A would be liable. For companies it's then a business decision whether the risks associated with collecting personal information are worth the costs.
“Even the best AI moderation has some error rate,” Guo said. He said the company’s models show that one to two posts out of every 10,000 viewed by the AI should have been caught on Parler but aren’t.
The problem is that AI moderation has two error rates, not one. The one he describes is the false-negative rate, the rate at which it mistakes a fail for a pass. The other that he doesn't mention is the false-positive rate, the rate at which it mistakes a pass for a fail. The two tend to go hand-in-hand: to reduce the false-negative rate you usually have to increase the false-positive rate and vice-versa. That's why you'll often have your doctor have you get a medical test and when the results come back they'll tell you they need you to come in for another test. The first one is a screening test that's fast and cheap and has a low false-negative rate, so if it comes back clean you're definitely clean. If it comes back positive they need to run a more sensitive (hence more expensive and time-consuming) test to get the real results, helped by the population for it having a much higher percentage of actual positives.
At least in the process Pai managed to botch it badly enough to give the Biden administration the perfect basis for rolling back Pai's policy changes related to network neutrality and restart the evaluation process from scratch.
On the post: Investigation Of ShotSpotter's Practices Is Raising Questions The Company's Angry Statement Really Doesn't Answer
I think their angry statement does answer the questions very clearly, the answer just isn't the one ShotSpotter wants to be able to give.
On the post: FTC Tries Tries Again With An Antitrust Case Against Facebook
Re:
The thing is, there's forced and then there's forced. You can be forced to do something even if nobody's holding a gun to your head. See "Hobson's choice": you can choose the first horse in line, or you can go without a horse. You aren't forced to choose that particular horse, but if you want a horse you don't have any other choices. You could go to another stable, assuming that a) there's another stable in town and b) it still has any horses left. If I have the biggest stable in town and any other stables are stuck with having only a couple of horses because I'm buying up all but a fraction of the feed, you'll likely find yourself without any other options. And is it really a choice if your only other choice is "do without"?
That's where Facebook is. When they weren't a monopoly and were facing competition, they went and used their money to buy up enough competitors that a critical mass of users wound up on their platform like it or not. Once that critical mass was reached, anyone looking at choices in social-media platforms was faced with the realization that they really didn't have a choice. The majority of their friends were on Facebook, so their choices were to use Facebook or to not interact with their friends on social media. Hobson's choice.
On the post: Netflix's Ramped Up War On VPNs Comes With Collateral Damage
My guess would be that the "collateral damage" is related to IPv4 address-space exhaustion. VPN providers discard "dirty" blocks of address space (ones flagged as being associated with VPNs and thus blacklisted) and move to new "clean" blocks. Used to be that those discarded blocks would take quite a while to be reallocated to a new owner, but now with IPv4 space being scarce they're being reallocated to new owners soon after being made available. The new owners now find themselves with address space that's on various blacklists because of the previous owners, but with no idea about the situation because they don't know who the previous owner even was let alone who's got those netblocks on blacklists.
The only fix there would lie with Netflix, not the ISPs or whoever got the "dirty" address space, because it's Netflix's blacklist and they're the only ones who can update it. They'd have to monitor address space releases and reallocations and remove stale entries from their blacklist, and I doubt they want to make that effort as long as they can get away with pointing fingers at someone else.
On the post: Netflix's Ramped Up War On VPNs Comes With Collateral Damage
Re:
That'd be my reaction too: "Why would my ISP know anything about why you guys added an address block to your blacklist, or even when you added it to know if they even owned that block at the time?".
On the post: Biden Warns That The Next Kinetic War Will Be The Result Of A Cyberattack, Which Is Stupid
I actually think Biden's right, the trigger that sets things in motion for the next major war will be some sort of cyber-attack. Not because the US or any other power intends to start a shooting war over it, but just because that event started a chain of misunderstandings and judgemental errors that will result in the shooting starting. Remember that the last time we were on the verge of a nuclear war (1995, Boris Yeltsin had actually retrieved the launch codes and was prepared to issue launch orders) it was because of the launch of a Norwegian research rocket. And Russia had even been told about the launch beforehand, the information simply hadn't been passed up the chain of command. My guess is it'll be the same thing: a cyber-attack (or more than one) gets everyone finger-pointing and causes disruptions that could be taken advantage of, then something else happens on one side that gets misinterpreted by the other side, someone in the chain decides that if they wait long enough for solid confirmation it'll be too late to act, or a couple of opposing pilots get a little too rambunctious and an accidental missile release triggers itchy trigger fingers on both sides, or an innocent vehicle gets misidentified as a threat and shot which triggers retaliation...
On the post: Supreme Court Says Patent Review Judges Are Unconstitutional, But It Can Be Fixed If USPTO Director Can Overrule Their Decisions
Another option would be to simply exercise Congress's authority under USC Article 3 section 1 to establish courts other than the Supreme Court. Make APJ judges actual judges, their proceedings actual court proceedings, and just to mess with the Justices make those courts directly beneath the Supreme Court itself so any appeals of their judgements go straight to the Justices for consideration. I don't even see anything in the Constitution that requires Congress to require confirmation of judges when establishing the courts, since they wouldn't be officers of the executive branch but be part of the judicial branch even if appointed by the director of an executive agency.
It seems like the Supremes are determined to eliminate any non-political or insulated-from-politics positions in the executive branch and return it completely to the spoils system.
On the post: DOJ Seizes Iranian News Org Websites; Raising Many Questions
Re: Re:
Why only limited to .ir domains? The US didn't limit itself to only seizing domains hosted in .us after all.
On the post: DOJ Seizes Iranian News Org Websites; Raising Many Questions
The question the US government needs to ask themselves is this: can the government of Iran seize the domains of US news outlets to prevent them from spreading what the Iranian government considers disinformation? I mean, the Iranian authorities can get the same kind of court order the US government got and all, so why would their seizure be any less legitimate?
On the post: Cop Who Led Strike Team Into Wrong House During Drug Raid Granted Immunity By Eleventh Circuit
One wonders what's going to happen one of these days in a state with a "stand your ground" law when someone notices armed intruders (the SWAT team) on a doorcam and proceeds to deal with said intruders as the law allows for. Sure they have body armor, but .30-caliber rounds from a hunting rifle will penetrate that easily and single-slug deer-hunting rounds from a shotgun won't even need to penetrate.
On the post: New 'Guardians Of The Galaxy' Game Has Game Streamers Worried Over Integral Music In The Game
Personally I think things like this are a good thing. Making the whole copyright mess cause actual financial damage to large corporations (by eg. having a triple-A mainstream game title flop because gamers can't safely promote it to others) may be the only way to get the mess fixed.
On the post: High School Responds To Student's Prank By Asking Local Law Enforcement To Step In And Investigate
Re: But think of the children
And of course the question of whether the web site in question even had any authentication set up, or if it (like so many) simply assumed nobody would ever use anything other than their real name to submit a quote. As one judge has pointed out, "not desired" doesn't equate to "unauthorized".
On the post: Content Moderation Case Study: Pretty Much Every Platform Overreacts To Content Removal Stimuli (2015)
And this is going to remain the case as long as platforms are liable for failing to remove content but not liable for wrongfully removing content, and copyright holders are entitled to damages for proven infringement but not liable for damages for wrongful claims of infringement.
One addition I'd love to make to the DMCA is that if a counter-notice is filed and the copyright holder fails to file an infringement suit within 90 days of being informed of the counter-notice, they are barred from ever bringing any further claims for the items involved (including future DMCA notices) and are automatically liable for statutory damages equal to the damages that would have resulted from their claimed infringement.
On the post: Why The Ninth Circuit's Decision In Lemmon V. Snap Is Wrong On Section 230 And Bad For Online Speech
Re:
By that logic the camera on the cel phone used should render the phone manufacturer liable. After all, it was the phone manufacturer, not the user, that put the camera in there and made the phone. And surely they knew that, since that camera can be used to document how fast you're travelling, teenagers would use it to document themselves going for exactly those same high speed numbers to show off just as you say. Similarly for the car manufacturers, knowing that simply putting a speedometer in the car would encourage reckless teenagers to capture an image of it to demonstrate just how fast they were capable of going.
I don't buy that.
On the post: Does Taking Down Content Lead Ignorant People To Believe It's More Likely To Be True?
I prefer the methods outlined in "David's Sling" in the data duels. Rather than removing the bad information, it's tagged with details of it's problems, links to supporting sources for correct information and the results of applying the corrected information to the argument the bad information was supposed to be in support of. This allows a response to the "think/do-the-research for yourself" line from the conspiracy theorists: "We did, and here's what we came up with.".
On the post: Citizen -- The App That Wants To Be A Cop -- Offered A $30,000 Bounty For The Apprehension Of An Innocent Person
Citizen, a division of Omni Consumer Products, Detroit, MI.
On the post: High School Teacher's Copyright Suit Against Netflix Gets Dismissed Because Coincidence Isn't Protectable
I really think the courts need to rethink their standards for sanctions. It's one thing when the defendant has to bring up things not mentioned in the complaint to get the complaint dismissed. But when the defense can win a motion for dismissal based solely on what's in the complaint, not sanctioning the plaintiff's attorney should be the exceptional case.
At the motion-to-dismiss stage the judge can rule only on matters of law requiring no determination of facts, and any ambiguity has to be taken in the light most favorable to the plaintiff. Given that, if the judge can find from the plaintiff's own complaint that the complaint doesn't state a case then the plaintiff's attorney certainly should have been able to come to the same conclusion. So either they filed it knowing it was baseless (which they're required not to do) or they filed it without doing their research on what the law and precedent say about it (which is also an ethical no-no) or they need to go back to law school. If it's really a valid complaint then they should be able to find at least one case that's still good law with a set of facts close enough to theirs to require a determination of facts to decide, which would let the complaint survive the motion to dismiss.
On the post: Rep. Zoe Lofgren Recognizes The Nuances And Complexities Of Regulating Internet Companies
The biggest problem with privacy issues is that the Internet companies tend to give you Hobson's choice: you either accept them doing whatever they want or you don't use their service at all, even when it's entirely possible for the service to work fine without collecting all that user data. Mostly that happens because they've settled on a business model that depends on selling that data for their revenue, without that they have no other source of revenue.
I don't like the idea of legislating business models, it tends to be too easy to abuse that power. For these cases I don't think it's necessary though. Just legislate liability: any company that collects personal information is 100% liable for the actual costs or a statutory damage amount, whichever is greater, plus legal costs and fees for every person whose data is compromised where that compromise can be traced back to their collection of the data. That means that if company A collects data and sells it to a data broker and that data broker's systems get compromised the data broker would be liable <i>and</i> company A would be liable. For companies it's then a business decision whether the risks associated with collecting personal information are worth the costs.
On the post: Parler Was Allowed Back In The Apple App Store Because It Will Block 'Hate Speech,' But Only When Viewed Through Apple Devices
The problem is that AI moderation has two error rates, not one. The one he describes is the false-negative rate, the rate at which it mistakes a fail for a pass. The other that he doesn't mention is the false-positive rate, the rate at which it mistakes a pass for a fail. The two tend to go hand-in-hand: to reduce the false-negative rate you usually have to increase the false-positive rate and vice-versa. That's why you'll often have your doctor have you get a medical test and when the results come back they'll tell you they need you to come in for another test. The first one is a screening test that's fast and cheap and has a low false-negative rate, so if it comes back clean you're definitely clean. If it comes back positive they need to run a more sensitive (hence more expensive and time-consuming) test to get the real results, helped by the population for it having a much higher percentage of actual positives.
On the post: Twitch Yanks Advertising Revenue From Popular 'Hot Tub Streamer' With No Warning Or Dialogue
Time for streamers to type "alternatives to Twitch" into a Google search and go researching.
On the post: Biden Revokes Trump's Silly Executive Order On Section 230; But It Already Did Its Damage
At least in the process Pai managed to botch it badly enough to give the Biden administration the perfect basis for rolling back Pai's policy changes related to network neutrality and restart the evaluation process from scratch.
Next >>