Were you going to vote for the guy who later endorsed the corporate-owned neoconservative warhawk? Because, given that, it seems like his principles weren't worth very much.
Such a bad and tired argument. He was doing it to try and stop Trump from getting elected. As a hardcore Sanders supporter, I am behind that decision 100%.
I fail to see how stopping Trump from winning is a palatable outcome if the alternate option is Clinton winning. She is diametrically opposed to just about every progressive position, I don't see how any progressive could support her in good conscience.
Unless of course it's more important that the team you identify with wins than what the principles that team stands for are.
If you truly consider yourself a progressive, take a minute to really think rationally about the situation. Put aside your emotions for Sanders. I know it's hard, and it took me a long time to come to grips with the situation before I could do that. I was very close to voting for Trump because of how terrible I thought Clinton was.
But, if you really think rationally about it, supporting Clinton is clearly the best option.
Under Clinton you still get many progressive policy pushes - paid leave, higher minimum wage, protecting abortion, stopping Republican voter suppression, and most importantly, SUPREME COURT NOMINATIONS! The supreme court nominations alone should be enough to convince you, and the rest is just icing on the cake. Not the mention, Trump supporters feel like they can be openly racist (which was my fear if Trump won, and sure enough, it has become a reality).
So yes, Clinton is a great option when compared to Trump.
> The supreme court nominations alone should be enough to convince you
They did. I do not want Hillary Clinton appointing judges to the supreme court any more than I want Trump doing it. Again, Hillary is not a progressive, so I see no reason to believe that her justice picks would be.
You haven't really made a convincing argument in her favor, you've essentially said "yes, she would have been an awful authoritarian who would continue and amplify the bad policies we already had, but at least she paid lip service to some inconsequential social wedge issues that serve as uniforms for my team!"
edit: It's especially disappointing that Hillary didn't lose by an even wider margin. Perhaps if the Democrats got bitten hard enough by their failed strategy we could have a chance at some real reform in four years. Instead people like you rewarded their avarice and deceit, so we can expect things to continue like this.
When she was a Senator, she did absolutely nothing on any of those fronts (or immigration)... even when Democrats controlled the White House and Congress, including a filibuster-proof Senate majority.
His only other alternative would have been to run on the Green ticket, and that probably wouldn't have worked out well for a number of reasons. As things shook out, he somehow ended up the de facto leader of a Democratic party that he hasn't even officially joined yet. If things had come out the other way, he would have pull for having brought his followers around and the chairmanship of a Senate committee. If he had been a failed Green party candidate, he would have been out of politics by now.
As it was, his answers to specific questions about what he endorsed about Clinton were so carefully phrased that he sounded like Clinton talking about emails.
But we were told that we couldn't vote for him because he was unelectable. We had to vote for the candidate who had been hated for decades, and who coincidentally directly or indirectly employed at least half of the Democratic party apparatus.
No matter, though. I'm sure that President Trump will be a wise and careful steward of his new surveillance network.
We've already asked you to please comment substantively, and there's nothing here besides inflammatory allegory. So we have to ask again: please don't post if you have nothing to say.
Not those of us who've been buying guns, but primarily rifles, of military utility in ever greater numbers, starting in the George W. Bush administration....
And let me quote Alexander Solzhenitsyn in The GULAG Archipelago on the general question of "whistling" vs. resisting:
And how we burned in the camps later, thinking: What would things have been like if every security operative, when he went out at night to make an arrest, had been uncertain whether he would return alive and had to say good-bye to his family?
Or if, during periods of mass arrests, as for example in Leningrad, when they arrested a quarter of the entire city, people had not simply sat there in their lairs, paling with terror at every bang of the downstairs door and at every step on the staircase, but had understood they had nothing left to lose and had boldly set up in the downstairs hall an ambush of half a dozen people with axes, hammers, pokers, or whatever else was at hand?
After all, you knew ahead of time that those bluecaps were out at night for no good purpose. And you could be sure ahead of time that you’d be cracking the skull of a cutthroat. Or what about the Black Maria [Government limo] sitting out there on the street with one lonely chauffeur — what if it had been driven off or its tires spiked.
The Organs would very quickly have suffered a shortage of officers and transport and, notwithstanding all of Stalin’s thirst, the cursed machine would have ground to a halt!
"Not those of us who've been buying guns, but primarily rifles, of military utility in ever greater numbers, starting in the George W. Bush administration"
Actually, gun ownership surged during the early Clinton administration in response to his party's push for gun bans and the "Brady Bill".
I can well believe that, but as I remember the big surge started after 9/11, when we were told the only thing we could personally do was to shop (support the already crashing economy), i.e. that we were are our own.
Quibble: our best numbers come from NICS checks, and they only started in 1998 based on 1993's Brady Bill. E.g. they capture sales of used guns, which the ATF otherwise doesn't have visibility into like they do with new stuff through Pittman Robertson (https://en.wikipedia.org/wiki/Pittman%E2%80%93Robertson_Fede...). And even that data is getting iffier as more and more people get concealed carry licenses, which generally allow you to skip the NICS check.
But, yeah, pretty much every effort, mooted or actual, at gun control prompts us ornery gun owners to resist, generally by buying some of what's under fire, so to speak.
I don't understand how these things get revealed in court. The cops are lying to secure warrants and charging documents, so why don't they keep lying when they're on the stand during a trial?
They're also under oath when they raise their right hand before a magistrate or other court officer and swear that the narratives they've written to support the warrant they're requesting is 100% true.
It's generally understood that police will lie like rugs at every point in the process, because the smooth functioning of the system requires it. But, when they're directly lying to a judge, the judge tends to take it personally (since no one likes being lied to) and has the power to impose immediate consequences.
Make no mistake, they do lie to judges, or at minimum give incantations with no regard to their truth value ("I noticed a furtive movement"). But easily-disproven lies are riskier.
a possible problem is when the software gets better then humans ... And you get many people who all can be identified by the vitness, witch one do you prosecute ?
on the other side of the coin, it's also a problem when the software is buggy. If innocent, it's tough to mount an effective defense without access to the algorithm.
You prosecute the one who seems most likely to have done it - who has a record, or was in the wrong place at the wrong time, or... Then you claim in court that there was less than a 1 in a million chance that he matched the suspect's photo, omitting the fact that there were numerous other matches. This is already happening with some of the less accurate forms of DNA testing.
“Better” here is problematic on many levels, not the least of which is an entirely automated (and likely privatized) legal system, streamlined for razor sharp "efficiency".
Every law an ethical implication, written by a culture frozen in a frail moment of history or worse, by the cultural detritus of corporate ownership, malignant power, or shady backrooms. Every line of code written by a human, with bias and imbalance.
The "efficiency" of such a system is carved out of the bedrock of dystopian science fiction.
> a possible problem is when the software gets better then humans ... And you get many people who all can be identified by the vitness, witch one do you prosecute ?
The non-rhetorical answer, according to how the system is supposed to work, is none of them. It's not possible for four independent actors to each have a 95% chance of being the perpetrator. If one has a 95% probability then the combined probability of all the others has to be <5%. And if the probability of the "most likely suspect" to have done it is e.g. 40%, that clearly isn't "beyond a reasonable doubt."
But this was all true even before computers. The system doesn't care about the math. A crime is committed, someone is to be punished, they put the most likely suspect in front of a jury and most of the time the jury will convict regardless.
In reality most of the time the accused will plead guilty to take a six month sentence over the 90+% chance of getting five years from a jury who doesn't understand (and is not given an opportunity to understand) how the probabilities actually work. And that incentive to plead guilty for a shorter sentence and lower legal bills is the same whether you're guilty or innocent.
Effectively, taking away your right to a trial by a jury of your peers. Plea bargains are one of the many elements of our legal system that is completely antithetical to actual justice and the Bill of Rights.
Kathy ONeill has written a book covering some of this new area of concern (hers is the use of poorly controlled algorithms in making life altering decisions (which teachers to fire, how long a prison sentence to recommend))
The existence of algorithms intending to help head teachers run their schools, or here in technology intended to help police catch criminals, the existence is not the problem.
It's the poor use, the lack of transparency and the lack of oversight
I expect cameras covering public spaces to be ubiquitous in a decade - but the difference between crime free neighbourhood and oppressive police state is who has access to the video feed, what algorithms they use to process the faces and so on.
This seems a sneaky move by the sheriffs department to avoid negative headlines. A bad PR move one suspects but the principle ... that's what we need to spend the next decade debating publically
> It's the poor use, the lack of transparency and the lack of oversight.
And the fact as (many of us are) programmers we know just how buggy, generally unreliable and flaky software can be.
Much like my instinctive recoil from IoT - "You want to put lowest cost manufactured network devices that have access to tonnes of personal data on my internal network and connect them to the internet..!", I find this trend deeply worrying, people who aren't techies tend to put undue weight on computer systems that usually give the correct results but can fail in subtle and not subtle ways with "Well it worked last time".
I've had people tell me "Well that's not what the computer says" when I absolutely knew that what the computer says is wrong, most recently that I'd submitted electricity meter readings (I had a screenshot of both the submission and the corresponding "Your meter reading was recieved" success message, not my first rodeo..).
To quote Jurassic Park "You where so busy figuring out if you could no one stopped to consider if they should".
It is interesting to consider that we may not actually want a 'crime-free' neighborhood, depending on how crimes are defined, and the distribution of personal and public costs associated with policing behavior and punishing offenders.
I quite agree. Facial recognition could have spotted Rosa Parks and simply shut the door to the bus.
Again, how we use the technology is vital. But open and free technologies and data feeds are far far less likely to be abused or if they are are much easier to spot.
>Using the technology, police can insert people with no criminal histories into virtual lineups without their knowledge
"This tool is too important to give up for the fight against 'bad guys'. Fears that it will spread are overblown. This technology will only be used to catch the worst of the worst" This is beyond Orwellian. Now we're all stock images for police lineups. I'd argue for royalties, but I'd settle for at least being notified.
What happens if someone picks your mugshot when it was included to just fill out the lineup? Do you now have to mount a costly legal while not having access to any details on the program?
I dont agree with its secret use for national security purposes, but at least i can understand it. Now its available to medium size cities with absolutely no oversight or transparency. This is the type of tool that allows unconstitutional investigations to be covered up with parallel construction.
On another note, Peter Thiel (Palantir) might have become the most powerful man in the world overnight.
Well, now not only is he the defacto head of the NSA, he has a direct line to the President, AND he'll most likely be Trump's go to guy on policies towards the Silicon Valley.
I know this has been coming for a while, but without the Trump win, he's not nearly as powerful.
This is greatly exaggerated. You are talking about Peter Thiel who is one of the best chess players of the country. He has strategical thinking and strong business focus, but he has almost nothing to do with the NSA. Palantir's clients are almost all of the agencies of the US working in the national security space not only NSA. If this is what you meant by that.
Well, several things that Trump has said in public during his presidential campaign can arguably be called "hate speech".
The UK legal definition is
"""
A person who uses threatening, abusive or insulting words or behaviour, or displays any written material which is threatening, abusive or insulting, is guilty of an offence if
a) he intends thereby to stir up racial hatred, or
(b) having regard to all the circumstances, racial hatred is likely to be stirred up thereby.
"""
Would you seriously suggest that when Trump says
"""
When Mexico sends its people, they're not sending their best. They're sending people that have lots of problems, and they're bringing those problems with us. They're bringing drugs. They're bringing crime. They're rapists.
"""
that this is does not fit the above definition of hate speech?
Who cares what the UK legal definition is? That definition is pretty arbitrary, and could include a lot of things that the average person wouldn't consider "hateful". The UK in particular has notoriously strict limits on freedom of speech that are incredibly antithetical to the US.
I don't really agree with the quote. But I think it's perfectly defensible. For starters, "illegal immigrant" isn't a race. Trump never said anything about hispanics in general, or people that were born in the US. I believe in context the quote didn't apply to legal immigrants either.
It's really disturbing to me that this climate has become so polarized, that anyone that says something you disagree with is "hate speech". Whether or not illegal immigration actually increases crime, it's a perfectly debatable issue. I think it's probably not true, but I can certainly understand why people would be concerned about it. I wouldn't consider those people racists.
In retrospect Thiel's bet on Trump was very good from a game theoretical perspective. Suppose he had supported Hillary and she won - he would have been one of very many Silicon Valley donors/supporters. Where as at the moment his position is almost unique and a lot more meaningful.
While Thiel might not be serially vindictive like Trump in terms of settling scores, he sure wasted no expense destroying Gawker.
I wonder how those who cut ties with him after the donation are feeling.
People don't understand probabilities. They don't accept that if someone has a 75% chance of being guilty, they may very well not be.
It happened again with FiveThirtyEight when everybody accused their "70% chance of Mrs Clinton winning" of being wrong, when it doesn't preclude Mr Trump winning.
> They don't accept that if someone has a 75% chance of being guilty, they may very well not be.
Nothing drove this point home to me more than X-COM: EU. Missing 3 successive 90% shots at the worst possible time would cause me to fling my keyboard across the room, but eventually, I learned to accept 90% isn't 100% :).
Nah, it's just how these games are. The first two fire emblem games on GameCube/wii were merciless in the same way on the normal modes. 80% hit chance sounds good until you miss and the entire battle goes south because of it. Since these games had permadeath, a miss could mean restarting if you didn't just hit reset mod battle.
Sometimes those odds are just those odds. A lot of gaming which uses pseudorandom had given a lot of people the gamblers fallacy as their understanding of probability. In some games it perhaps makes sense; in league of legends I believe the Critical hit chance is pseudo random to feel less swingy, which is probably healthier for a competitive game. Diablo 3 weights it's unique drop rates to ensure that players get a steady supply of relevant uniques since there is no trade economy in d3 anymore. This also makes sense in isolation.
Other games though that don't have as strong of a reason to fudge the roll give a really bad impression of random to players.
You might consider using a term other than pseudo random for that, since pseudo random number generators generally try to be uniformly random, as opposed to the sort of thing you're talking about.
Interesting, thanks for making that point. I think it might be best to refer to one as a prng and the other as a prd (generator vs distribution) although implementing a prd creates a prng, some prng attempt to target a real random distribution.
Late reply but thanks for the heads up - I wasn't aware that prng was a term that I could be using as opposed to prd. The above comment is accurate in how I first came across the term as it's pretty commonly used in a lot of games in a lay person's fashion.
Because games are created for humans, and humans are bad at understanding probabilities. So when 90% shot fails three times in a row, overwhelming majority of players will think that the game is cheating.
While you learned a good lesson, hit percentages in games are very often lies.
I'm seeing conflicting data for X-COM: EU specifically. Some say it's fair (except for cheating in your favor when you're losing on easier modes), but this guy collected a quite-significant number of samples and got results that looked like the square of the displayed hit chance: https://steamcommunity.com/app/200510/discussions/0/61393794...
The 70% probably was wrong, in the sense of being uncalibrated. 538 did their best to fit the data to past elections, but there haven't really been enough elections to fit a good model.
> And when I see Dan, Nate Silver and various other insider pundits saying what they are saying, I have been to enough rodeos to know that thoughts are being planted from the Obama-Clinton consultant class (not you, and you know who I mean). This is not helpful to Hillary, quite the contrary, Bernie has zero chance of being nominated and Hillary is going to need every liberal supporting her to vote, which is why I am being very friendly to Bernie and will have some credibility when the time comes to encourage his people to get on the train with Hillary.
There's no question that 538 is anti-Trump. That shouldn't affect their statistical model, which gave much higher odds to Trump than most other models. I don't see the relevance of either of those links.
The point of the other links is that Nate is compromising his professional image through punditry.
If I asked a consultant for advice on the future of my sector of the economy and then found a series of blog posts bashing my industry which were all unfounded, I would call his judgment into question.
At best, he is trying to influence the very thing he is supposed to me impartially analysing.
> What happens if someone picks your mugshot when it was included to just fill out the lineup?
I don't see how it would be any different than traditional lineups, where if a witness selects one of the foils, they've just failed to select the police suspect. Police don't just stack a lineup with potential suspects in order to decide who to investigate.
Would we have problem if he was identified by manual search in the database? If not, I don't see why using technology to remove the manual labor is worse.
Is the problem in the existence of the database? Then I would not make this a discussion about facial recognition software but about mugshot databases instead.
> f not, I don't see why using technology to remove the manual labor is worse.
What's its accuracy? How was the surveillance obtained? Is the time of the photo accurate or was the server clock set incorrectly? Was it obtained constitutionally or through an illegal investigation? What if it's incorrect, can I have the algorithm to help mount a legal defense
I suppose that the result is still evaluated by a human so the question about accuracy is irrelevant. All of the other questions can be asked irrespective if pattern recognition or manual labor was used.
It shouldn't really matter what the accuracy of the software is being used properly. The facial recognition software should be used as a tool to help narrow down potential suspects. It shouldn't be the judge jury and executioner to decide if someone was indeed at the scene of a crime. The officers using the system need to be properly trained to understand the limitations of the system. For example just because the system finds a close match, it is only searching a database of people it already knows about. The actual suspect may not be in the system. Traditional police detective work still applies. The face match is just a lead not hard evidence.
Also, the systems use should be public (just like the police's use of fingerprint biometrics is). Problems come in when you try to use this or any technology to automate law enforcement. Human judgement and police work need to always be part of the equation.
Facial recognition should be thought of as a username not a username and password combination. It shouldn't be the only piece of evidence in a case to prosecute.
> It shouldn't really matter what the accuracy of the software is being used properly.
You seem to be under the impression that our legal system is efficient. It is not at all like that. You could be falsely arrested. It could take months before you actually able to have your case in front of a judge. In the meantime, you have to bail out of jail and hire a lawyer. That's thousands if not tens of thousands.
In the end, the prosecution can just drop the case with any consequence and you've lost time, money, and possibly your job.
Scale matters. The fact that I'm on the street is public knowledge. But yes, I'd have a problem with an automatically searchable database of all my moves in public places recorded by cameras.
Edit: manually searchable footage is in-between. Manual search has natural throttling which makes it a bit less invasive against privacy.
Anyone would. You don't have a right to privacy in public, but it's another thing for ubiquitous surveillance to track and detail all of your movements, and then save them indefinitely. Not having a right to privacy in public does not mean it's okay to have the equivalent of my google maps location history.
The problem is that with a large enough database, no matter who's actually in your input photo, you're going to find one or more matches. And if your matching algorithm is any good, they're gonna look really really similar to your input photo. So you've got a pre-screened list of suspects that you can pick from, to put in front of a jury and say "see, here's a photo of this guy doing the crime, the computer says it's him."
The answer is that false positives are a thing, and a test sufficiently specific to run a dozen times probably doesn't scale to a million times.
Similarly, you can hit phase transitions in capability where more of the same isn't just the same any more. That's why we let police look at things by eye from public space, but count it as a search if they use technology to enhance their view (such as IR).
The main concern is that technology is allowing police to expend more policing effort than we want them to. Their powers aren't supposed to be perfect, and pushing their success rate too high is dangerous for society.
Police would get their human staff cut if they sat them down to build a cross correlation database of citizen activity out of photos from all around town. They'd be called the Stasi and worse.
Why is it any different because they use computers instead of humans?
Ed: I think of modern policing in the US as a sort of autoimmune disease, prompted by a combination of natural proclivity and particularly nasty, if brief, shock to the system a decade-and-a-half ago. Policing in the abstract is good for society; policing like we are now is not.
The main concern is that technology is allowing police to expend more policing effort than we want them to. Their powers aren't supposed to be perfect, and pushing their success rate too high is dangerous for society.
Exactly. Imagine if we had perfect enforcement of laws against sodomy or miscegenation. In addition to the increase in human misery, it would have been impossible for a reservoir of free personally-invested persons to circulate in society, agitating for change.
Manual search forces the police to expend energy enforcing the laws, which forces them to prioritize which laws to expend energy enforcing, which (sometimes, when we're lucky) forces them to be responsive to broader societal values.
These databases create a vision of the future which is the very same boot stomping on the very same face in the very same quantity, forever.
This article reminds me of the one where the guy in prison discovers the existence of the Stingray devices. [1] A fine example of research and investigative work to uncover a serious violation of rights.
I'm in that database because my photo was taken for a driver's license. Imagine the system is not 100% reliable (which of course it isn't), I wind up being matched as a probable, and become a target for investigation, and potentially arrested. At this point I am only guilty of driving a car and will have to defend myself by hiring a lawyer. Eventually it will be obvious I couldn't be the person. By then my life is mostly in ruin. Yet there is nothing I can do to avoid this possibility legally.
In the USA that doesn't help. If you don't have a driver's license, you have to have a "state-issued ID card", which is the same except it doesn't confer the right to drive. The only way to avoid both is to opt out of society completely, which is effectively giving up living in this country.
They're expensive, most Americans don't have them, especially the poorer, and they're a greater hassle: people expect a driver's license. And once you've had one, they have your photo; too late to change.
Also, not having a driver's license is looked down on in most of the USA, outside a few major cities. It constrains job/dating/social/living location opportunities.
In Washington I was able to request to not have a photo on my license which I started doing when they started using a facial database. The license says "NOT VALID FOR IDENTIFICATION" in large red letters, so I can't use it to buy beer. But I can use it to drive a car which is all I really need it for.
Just learned that California, Missouri, Louisiana, Mississippi, Maine, New Hampshire, and Vermont are the lone states that do not use facial recognition tech on DMV photos.
I imagine though that eventually the FBI, State Dept, states, etc. will all be combined
Or, you know, maybe our founders knew a good idea when they adopted this federalism thing.
As someone born, raised in and retired to the great state of Missouri, which is also the sole state that doesn't play big brother with narcotic prescriptions, I rather like the idea of imagining that we won't combine our law enforcement agencies in such a vertical manner.
The problem of course is that those scores aren't very good either, because they don't distinguish between e.g. libertarian peaceniks, right-wing anarchists, and biker gang bangers. But the solution is just more data; cross-referencing with a consumer purchase database, for example. Eventually the police will get their responses well-calibrated. I'd assume this sort of thing is what Trump is referring to when he says he supports profiling.
Because at some level, policing is about delivering a product (a policeman) to a consumer (a victim / criminal in a bad situation), so standard marketing tactics (market segmentation, revenue management, etc.) apply. Many other businesses have access to these databases and use them to beneficial effect; there's no reason police shouldn't jump onto the big data bandwagon too to increase customer satisfaction.
I would imagine that this database is going to be used to supplement police work rather than substitute it. I don't think police or prosecutors will try to press charges based on the database searches alone because it is very unlikely that something like that would hold up in court. They will also have to show other evidence that the person in question did in fact commit the crime. For example, in this case, it seems like the software justified the drug sting, but I imagine the drugs and paraphernalia found at the scene was what justified the actual arrest and prosecution.
What will a neo-authoritarian President do with such power? How about one who threatens to use his legal power to suppress and punish his political enemies? Would law enforcement, such as the FBI, act on behalf of one party?
It is on every one of us to act. What is your company doing to create or reduce these risks? What is your legislator doing? If the United States is the land of the free, if that is more than words, then its citizens must act like it.
As many veterans of prior struggles have said: Don't get scared, don't get mad - Organize!
No, they generally do the latter, and generally well.
One problem is there just aren't enough killers etc. anymore, crime rates are way down from their peaks not so many decades ago. But, somehow, where this has happened, it hasn't resulted in a reduction in the size of what I like to call the police-judicial complex. This book's underlying thesis is that too avoid unthinkable layoffs, the complex needs a steady diet of "the clueless" to keep running at it's current tempo: Arrest-Proof Yourself (https://www.amazon.com/Arrest-Proof-Yourself-Dale-C-Carson/d...) I haven't read the 2nd edition I linked to, but I can't recommend the 1st highly enough in every respect, theory to practice.
> No, they generally do the latter, and generally well.
Minor quibble, but I don't find a 50% clearance rate on murders to be good. [0] I also don't put much faith in the accuracy of many (most) police departments numbers as they "Juke the Stats" [1]
Stranger murders are particularly difficult cases to clear; here I'm talking about "real felonies" in general, and I'm also speaking very generally. E.g. Chicago's murder clearance rate for 2016 murders has plummeted to 21% (http://www.chicagotribune.com/news/local/breaking/ct-chicago...).
Here in my above average crime rate (old mining town) corner of SW Missouri, the murder clearance rate is very high. However, the first relevant thing I found on Google was a State Police summary document for 2013 (https://www.mshp.dps.missouri.gov/MSHPWeb/SAC/pdf/Crime%20in...), violent crime clearance rate was ~50%, murder ~64%. And Missouri overall might be a good proxy for the US in general, we have everything from two old urban centers to the most rural of areas.
So, yeah, you could well be right in #0, and you're certainly right with #1.
everyone forgets the obvious: line ups don't work.
you got a broken concept, and throw expensive technology that only contribute to the illusion but doesn't address any of the flaws, just make it easier for the enforcers of the broken system. the result is people abusing even more the broken system, because now it's easier, and some tech provider getting very rich, and then sponsoring campaigns so that the system never change.
116 comments
[ 3.4 ms ] story [ 173 ms ] threadUnless of course it's more important that the team you identify with wins than what the principles that team stands for are.
But, if you really think rationally about it, supporting Clinton is clearly the best option.
Under Clinton you still get many progressive policy pushes - paid leave, higher minimum wage, protecting abortion, stopping Republican voter suppression, and most importantly, SUPREME COURT NOMINATIONS! The supreme court nominations alone should be enough to convince you, and the rest is just icing on the cake. Not the mention, Trump supporters feel like they can be openly racist (which was my fear if Trump won, and sure enough, it has become a reality).
So yes, Clinton is a great option when compared to Trump.
They did. I do not want Hillary Clinton appointing judges to the supreme court any more than I want Trump doing it. Again, Hillary is not a progressive, so I see no reason to believe that her justice picks would be.
You haven't really made a convincing argument in her favor, you've essentially said "yes, she would have been an awful authoritarian who would continue and amplify the bad policies we already had, but at least she paid lip service to some inconsequential social wedge issues that serve as uniforms for my team!"
edit: It's especially disappointing that Hillary didn't lose by an even wider margin. Perhaps if the Democrats got bitten hard enough by their failed strategy we could have a chance at some real reform in four years. Instead people like you rewarded their avarice and deceit, so we can expect things to continue like this.
When she was a Senator, she did absolutely nothing on any of those fronts (or immigration)... even when Democrats controlled the White House and Congress, including a filibuster-proof Senate majority.
The broader point is that she made no efforts in those areas on the legislative front, even when Democrats held the majority in both houses.
As it was, his answers to specific questions about what he endorsed about Clinton were so carefully phrased that he sounded like Clinton talking about emails.
No matter, though. I'm sure that President Trump will be a wise and careful steward of his new surveillance network.
And let me quote Alexander Solzhenitsyn in The GULAG Archipelago on the general question of "whistling" vs. resisting:
And how we burned in the camps later, thinking: What would things have been like if every security operative, when he went out at night to make an arrest, had been uncertain whether he would return alive and had to say good-bye to his family?
Or if, during periods of mass arrests, as for example in Leningrad, when they arrested a quarter of the entire city, people had not simply sat there in their lairs, paling with terror at every bang of the downstairs door and at every step on the staircase, but had understood they had nothing left to lose and had boldly set up in the downstairs hall an ambush of half a dozen people with axes, hammers, pokers, or whatever else was at hand?
After all, you knew ahead of time that those bluecaps were out at night for no good purpose. And you could be sure ahead of time that you’d be cracking the skull of a cutthroat. Or what about the Black Maria [Government limo] sitting out there on the street with one lonely chauffeur — what if it had been driven off or its tires spiked.
The Organs would very quickly have suffered a shortage of officers and transport and, notwithstanding all of Stalin’s thirst, the cursed machine would have ground to a halt!
Actually, gun ownership surged during the early Clinton administration in response to his party's push for gun bans and the "Brady Bill".
Quibble: our best numbers come from NICS checks, and they only started in 1998 based on 1993's Brady Bill. E.g. they capture sales of used guns, which the ATF otherwise doesn't have visibility into like they do with new stuff through Pittman Robertson (https://en.wikipedia.org/wiki/Pittman%E2%80%93Robertson_Fede...). And even that data is getting iffier as more and more people get concealed carry licenses, which generally allow you to skip the NICS check.
But, yeah, pretty much every effort, mooted or actual, at gun control prompts us ornery gun owners to resist, generally by buying some of what's under fire, so to speak.
Make no mistake, they do lie to judges, or at minimum give incantations with no regard to their truth value ("I noticed a furtive movement"). But easily-disproven lies are riskier.
Or when the software is not open source and therefor can contain elements of corruption because not being reviewed by the public.
Every law an ethical implication, written by a culture frozen in a frail moment of history or worse, by the cultural detritus of corporate ownership, malignant power, or shady backrooms. Every line of code written by a human, with bias and imbalance.
The "efficiency" of such a system is carved out of the bedrock of dystopian science fiction.
The non-rhetorical answer, according to how the system is supposed to work, is none of them. It's not possible for four independent actors to each have a 95% chance of being the perpetrator. If one has a 95% probability then the combined probability of all the others has to be <5%. And if the probability of the "most likely suspect" to have done it is e.g. 40%, that clearly isn't "beyond a reasonable doubt."
But this was all true even before computers. The system doesn't care about the math. A crime is committed, someone is to be punished, they put the most likely suspect in front of a jury and most of the time the jury will convict regardless.
In reality most of the time the accused will plead guilty to take a six month sentence over the 90+% chance of getting five years from a jury who doesn't understand (and is not given an opportunity to understand) how the probabilities actually work. And that incentive to plead guilty for a shorter sentence and lower legal bills is the same whether you're guilty or innocent.
The existence of algorithms intending to help head teachers run their schools, or here in technology intended to help police catch criminals, the existence is not the problem.
It's the poor use, the lack of transparency and the lack of oversight
I expect cameras covering public spaces to be ubiquitous in a decade - but the difference between crime free neighbourhood and oppressive police state is who has access to the video feed, what algorithms they use to process the faces and so on.
This seems a sneaky move by the sheriffs department to avoid negative headlines. A bad PR move one suspects but the principle ... that's what we need to spend the next decade debating publically
And the fact as (many of us are) programmers we know just how buggy, generally unreliable and flaky software can be.
Much like my instinctive recoil from IoT - "You want to put lowest cost manufactured network devices that have access to tonnes of personal data on my internal network and connect them to the internet..!", I find this trend deeply worrying, people who aren't techies tend to put undue weight on computer systems that usually give the correct results but can fail in subtle and not subtle ways with "Well it worked last time".
I've had people tell me "Well that's not what the computer says" when I absolutely knew that what the computer says is wrong, most recently that I'd submitted electricity meter readings (I had a screenshot of both the submission and the corresponding "Your meter reading was recieved" success message, not my first rodeo..).
To quote Jurassic Park "You where so busy figuring out if you could no one stopped to consider if they should".
Again, how we use the technology is vital. But open and free technologies and data feeds are far far less likely to be abused or if they are are much easier to spot.
"This tool is too important to give up for the fight against 'bad guys'. Fears that it will spread are overblown. This technology will only be used to catch the worst of the worst" This is beyond Orwellian. Now we're all stock images for police lineups. I'd argue for royalties, but I'd settle for at least being notified.
What happens if someone picks your mugshot when it was included to just fill out the lineup? Do you now have to mount a costly legal while not having access to any details on the program?
I dont agree with its secret use for national security purposes, but at least i can understand it. Now its available to medium size cities with absolutely no oversight or transparency. This is the type of tool that allows unconstitutional investigations to be covered up with parallel construction.
On another note, Peter Thiel (Palantir) might have become the most powerful man in the world overnight.
Overnight? Thiel has been building Palantir for a long time now. It's older than Twitter.
It sucks, but I'm not sure what can be done. The technology is only going to get better and cheaper and more accessible.
I know this has been coming for a while, but without the Trump win, he's not nearly as powerful.
not a fan of thiel, but this is not even close to true
Can we PLEASE stop with that nonsense? You might not agree. It doesn't mean it's "hate".
The election is over. Get over it. Please.
The UK legal definition is
"""
A person who uses threatening, abusive or insulting words or behaviour, or displays any written material which is threatening, abusive or insulting, is guilty of an offence if
a) he intends thereby to stir up racial hatred, or
(b) having regard to all the circumstances, racial hatred is likely to be stirred up thereby.
"""
Would you seriously suggest that when Trump says
"""
When Mexico sends its people, they're not sending their best. They're sending people that have lots of problems, and they're bringing those problems with us. They're bringing drugs. They're bringing crime. They're rapists.
"""
that this is does not fit the above definition of hate speech?
"hate speech" as opposed to "hate".
Please just stop with the campaign sound bite nonsense. I'm weary of it.
I don't really agree with the quote. But I think it's perfectly defensible. For starters, "illegal immigrant" isn't a race. Trump never said anything about hispanics in general, or people that were born in the US. I believe in context the quote didn't apply to legal immigrants either.
It's really disturbing to me that this climate has become so polarized, that anyone that says something you disagree with is "hate speech". Whether or not illegal immigration actually increases crime, it's a perfectly debatable issue. I think it's probably not true, but I can certainly understand why people would be concerned about it. I wouldn't consider those people racists.
I am reminded of this talk by Sid Meier: https://youtu.be/bY7aRJE-oOY?t=19m27s
People don't understand probabilities. They don't accept that if someone has a 75% chance of being guilty, they may very well not be.
It happened again with FiveThirtyEight when everybody accused their "70% chance of Mrs Clinton winning" of being wrong, when it doesn't preclude Mr Trump winning.
Nothing drove this point home to me more than X-COM: EU. Missing 3 successive 90% shots at the worst possible time would cause me to fling my keyboard across the room, but eventually, I learned to accept 90% isn't 100% :).
Sometimes those odds are just those odds. A lot of gaming which uses pseudorandom had given a lot of people the gamblers fallacy as their understanding of probability. In some games it perhaps makes sense; in league of legends I believe the Critical hit chance is pseudo random to feel less swingy, which is probably healthier for a competitive game. Diablo 3 weights it's unique drop rates to ensure that players get a steady supply of relevant uniques since there is no trade economy in d3 anymore. This also makes sense in isolation.
Other games though that don't have as strong of a reason to fudge the roll give a really bad impression of random to players.
For example, in DotA: http://wiki.teamliquid.net/dota2/Pseudo_Random_Distribution
http://web.archive.org/web/20141004060133/http://old.serenes...
I'm seeing conflicting data for X-COM: EU specifically. Some say it's fair (except for cheating in your favor when you're losing on easier modes), but this guy collected a quite-significant number of samples and got results that looked like the square of the displayed hit chance: https://steamcommunity.com/app/200510/discussions/0/61393794...
https://wikileaks.org/podesta-emails/emailid/28878
From:brentbbi@webtv.net
To: john.podesta@gmail.com
Date: 2015-07-09 09:44
Subject: CNN and bed-wetting
> And when I see Dan, Nate Silver and various other insider pundits saying what they are saying, I have been to enough rodeos to know that thoughts are being planted from the Obama-Clinton consultant class (not you, and you know who I mean). This is not helpful to Hillary, quite the contrary, Bernie has zero chance of being nominated and Hillary is going to need every liberal supporting her to vote, which is why I am being very friendly to Bernie and will have some credibility when the time comes to encourage his people to get on the train with Hillary.
Examples of his work
https://i.sli.mg/fYycbE.jpg
If I asked a consultant for advice on the future of my sector of the economy and then found a series of blog posts bashing my industry which were all unfounded, I would call his judgment into question.
At best, he is trying to influence the very thing he is supposed to me impartially analysing.
I don't see how it would be any different than traditional lineups, where if a witness selects one of the foils, they've just failed to select the police suspect. Police don't just stack a lineup with potential suspects in order to decide who to investigate.
Is the problem in the existence of the database? Then I would not make this a discussion about facial recognition software but about mugshot databases instead.
What's its accuracy? How was the surveillance obtained? Is the time of the photo accurate or was the server clock set incorrectly? Was it obtained constitutionally or through an illegal investigation? What if it's incorrect, can I have the algorithm to help mount a legal defense
Also, the systems use should be public (just like the police's use of fingerprint biometrics is). Problems come in when you try to use this or any technology to automate law enforcement. Human judgement and police work need to always be part of the equation.
Facial recognition should be thought of as a username not a username and password combination. It shouldn't be the only piece of evidence in a case to prosecute.
You seem to be under the impression that our legal system is efficient. It is not at all like that. You could be falsely arrested. It could take months before you actually able to have your case in front of a judge. In the meantime, you have to bail out of jail and hire a lawyer. That's thousands if not tens of thousands.
In the end, the prosecution can just drop the case with any consequence and you've lost time, money, and possibly your job.
Edit: manually searchable footage is in-between. Manual search has natural throttling which makes it a bit less invasive against privacy.
Not a joke. Jurors put a lot of faith in forensic techniques
Similarly, you can hit phase transitions in capability where more of the same isn't just the same any more. That's why we let police look at things by eye from public space, but count it as a search if they use technology to enhance their view (such as IR).
The main concern is that technology is allowing police to expend more policing effort than we want them to. Their powers aren't supposed to be perfect, and pushing their success rate too high is dangerous for society.
Police would get their human staff cut if they sat them down to build a cross correlation database of citizen activity out of photos from all around town. They'd be called the Stasi and worse.
Why is it any different because they use computers instead of humans?
Ed: I think of modern policing in the US as a sort of autoimmune disease, prompted by a combination of natural proclivity and particularly nasty, if brief, shock to the system a decade-and-a-half ago. Policing in the abstract is good for society; policing like we are now is not.
Exactly. Imagine if we had perfect enforcement of laws against sodomy or miscegenation. In addition to the increase in human misery, it would have been impossible for a reservoir of free personally-invested persons to circulate in society, agitating for change.
Manual search forces the police to expend energy enforcing the laws, which forces them to prioritize which laws to expend energy enforcing, which (sometimes, when we're lucky) forces them to be responsive to broader societal values.
These databases create a vision of the future which is the very same boot stomping on the very same face in the very same quantity, forever.
"Show me a man that has never broken the law and I shall show you a man that has never driven a car."
https://www.fantasticfiction.com/m/richard-miller/squed.htm
[1]. http://www.theverge.com/2016/1/13/10758380/stingray-surveill...
Don't drive? I know that sucks, but with self-driving cars on the horizon, these sort of "capitulation of rights" scenarios should not need to exist.
We are very vulnerable (in terms of rights) when driving in the US.
We are very vulnerable (in terms of rights) in the US.
Why? There are alternatives for ID and proof-of-age, such as a passport booklet/card.
Also, not having a driver's license is looked down on in most of the USA, outside a few major cities. It constrains job/dating/social/living location opportunities.
Most laws only require people requesting proof of identity accept an in-state drivers license or military ID.
I imagine though that eventually the FBI, State Dept, states, etc. will all be combined
As someone born, raised in and retired to the great state of Missouri, which is also the sole state that doesn't play big brother with narcotic prescriptions, I rather like the idea of imagining that we won't combine our law enforcement agencies in such a vertical manner.
[0] http://www.eonline.com/resize/300/300/akns-images.eonline.co...
The problem of course is that those scores aren't very good either, because they don't distinguish between e.g. libertarian peaceniks, right-wing anarchists, and biker gang bangers. But the solution is just more data; cross-referencing with a consumer purchase database, for example. Eventually the police will get their responses well-calibrated. I'd assume this sort of thing is what Trump is referring to when he says he supports profiling.
See e.g. http://www.citylab.com/crime/2015/10/uk-police-departments-d... http://www.theatlantic.com/national/archive/2015/02/more-pol...
It is on every one of us to act. What is your company doing to create or reduce these risks? What is your legislator doing? If the United States is the land of the free, if that is more than words, then its citizens must act like it.
As many veterans of prior struggles have said: Don't get scared, don't get mad - Organize!
One problem is there just aren't enough killers etc. anymore, crime rates are way down from their peaks not so many decades ago. But, somehow, where this has happened, it hasn't resulted in a reduction in the size of what I like to call the police-judicial complex. This book's underlying thesis is that too avoid unthinkable layoffs, the complex needs a steady diet of "the clueless" to keep running at it's current tempo: Arrest-Proof Yourself (https://www.amazon.com/Arrest-Proof-Yourself-Dale-C-Carson/d...) I haven't read the 2nd edition I linked to, but I can't recommend the 1st highly enough in every respect, theory to practice.
Minor quibble, but I don't find a 50% clearance rate on murders to be good. [0] I also don't put much faith in the accuracy of many (most) police departments numbers as they "Juke the Stats" [1]
[0] http://laist.com/2015/01/26/almost_half_of_the_murders_in_th...
[1] https://www.youtube.com/watch?v=xH_6_8NOfwI
Here in my above average crime rate (old mining town) corner of SW Missouri, the murder clearance rate is very high. However, the first relevant thing I found on Google was a State Police summary document for 2013 (https://www.mshp.dps.missouri.gov/MSHPWeb/SAC/pdf/Crime%20in...), violent crime clearance rate was ~50%, murder ~64%. And Missouri overall might be a good proxy for the US in general, we have everything from two old urban centers to the most rural of areas.
So, yeah, you could well be right in #0, and you're certainly right with #1.
you got a broken concept, and throw expensive technology that only contribute to the illusion but doesn't address any of the flaws, just make it easier for the enforcers of the broken system. the result is people abusing even more the broken system, because now it's easier, and some tech provider getting very rich, and then sponsoring campaigns so that the system never change.