I've been wondering about the UK contact tracing app because they seem to be deliberately misleading saying that data is secure on the phone, yet it's a centralised model, and they are using bullshit terms like "clinically secure algorithm" to describe the one-time codes;
Is the source code of these apps something that could be FOI requested from NHSx seeing as it is publicly funded by the tax payer?
I think this is actually quite innovative from the Uk government, take the existing NHS brand which has been a thorn in the side of the Conservatives for decades, co-opt the brand for dodgy spy software thta not only tarnishes the NHS brand, but also continues the UK government attempt at tracking every step every citizen takes.
I remember the desperate efforts of Theresa May to pass the snoopers charter for such a long time. And when she became a PM it was on the top of her list, and she did it. And now with that unethical filthy blob as a PM, I only expect the worse of him and his lackeys..
IMHO Apple and Google should have finished the job and "volunteer" to make the app as well as "donate" some of the storages for the data. Not that I trust US companies to respect anyone's privacy, or to the fact that a gag order stapled to a subpoena would give all that data to the "5 eyes"(now 14-15).
I like my health as much as the next living human. But there are so many governments that will jump on this opportunity that it makes me want to avoid this app.
> I remember the desperate efforts of Theresa May to pass the snoopers charter for such a long time. And when she became a PM it was on the top of her list, and she did it.
RIPA was passed in the year 2000 under a Labour government when Jack Straw was Home Secretary.
I don't think that's what they're doing. Intersection of people who care about privacy and people who wouldn't see through that tactic is tiny.
They are using the NHS brand to get high adoption.
Plus the UK govt slogan right now is literally "protect the NHS", the existing Conservative party opposition to the NHS is basically gone now, hopefully for good.
> Plus the UK govt slogan right now is literally "protect the NHS", the existing Conservative party opposition to the NHS is basically gone now, hopefully for good.
One does not follow from the other. Slogans substitute for competency and funding. The opposition is both ideological and directly financially motivated, so I expect the fragmentary privatization to continue.
Watch out for a "now we must pay for coronavirus" user surcharge appearing ...
Oh wow I had no idea Palintir was involved, I feel like I've finally turned in to the paranoid old fool that people laugh at, but this really upsets me to think of the millions of people who are going to install this without so much as a thought.
The thing about those paranoid old fools is that some of them are right. But where they are stuck at the level of trying to convince others to just "look with your eyes, see the truth right in front of you!" those with actual experience with praxis know it's much more fruitful to put your efforts toward political campaigns. Grassroots is a thing but it's always some weird combination of coercion and rhetoric about morals and ethics, because most people don't care so you have to make them care before the movement picks up steam.
Talk is cheap. Have vegetarians converted a lot of people to give up meat?
No, you need viable alternatives for people to switch. Like the Impossible Burger and so on. Until then talking won’t change anything. Not even Snowden level revelations would change it. Sure you can try to use government to fight government. Or... just build the alternative.
We built the Web. We killed AOL MSN and Compuserve.
Build open source, end to end encrypted, self healing and rebalancing networks that use a version of Kademlia DHT that removes IP addresses from each hop. And also run consensus in small groups about stuff.
That’s the future right there. Trouble is, we are only seeing its infancy. I can think of about 2-3 projects that are pulling it off. And they have been working for years.
PS: When this happens, user accounts and quotas will be replaced with crypto, and centralized servers and databases will be only by opt in. They wouldn’t automatically be a thing just because they provide the infrastructure. Infra should become completely commoditized.
PPS: But. You’ll have to worry massivelu about botnets, sybil attacks and massive disinfo campaigns and reputationao attacks by sleeper AI bots. “Fun” times ahead ...
I disagree with that analogy. Palantir isn’t getting them more data. It’s helping them use it more effectively. The organizations already have all the data recorded.
If you don’t like the data that organizations record, I’m right there with you, but demonizing the tools that allow them to access data more effectively is an absurd stance to take. Their system also does things like add an immutable audit log, or enforce requirements on how data can be accessed.
Working with law enforcement, before Palantir there was nothing preventing an officer from looking up their ex husband or wife, or celebrities, or anyone at random. With it, there’s an audit trail of every search, and most of them need a case number as justification. Even then, certain searches would automatically get flagged for review.
We can go back to giving people direct and unaudited SQL access if you’d like, but I’d prefer some more accountability, even if the downside is organizations being able to use data _they already have_ more effectively.
The analogy seems apt; weapons dealers don't give warlords armies (if they did, they'd be called mercenaries.) They give the armies of warlords the tools they need to be more effective. The violent acts, killing people or holding PII, are still performed by the warlord's organization, but does that excuse the people selling the tools used?
I guess we’ll just have to agree to disagree. I just don’t think that complaining about a tech company when what you actually dislike is what governments or other organizations are doing is going to impact anything.
But Palantir has been the “tech boogieman” since before I even left in 2014, so it probably always will be. It is admittedly easier to blame a smaller tech company as a scapegoat than try to address the larger governmental or organizational actors that you actually have grievances against.
Just know it won’t do anything. The people who complain online aren’t their target market and never will be. And the paranoia lends way more credence to their dressed up search engine than it really deserves on technical merit alone.
So in a way the people complaining about it are making the issue they’re so upset about worse.
> It is admittedly easier to blame a smaller tech company as a scapegoat than try to address the larger governmental or organizational actors that you actually have grievances against.
Or you can do both. You might as well suppose that people are critical of the mercenary company Blackwater because they want to avoid criticizing America foreign policy and military exploits. I don't think that's accurate at all, it's been my experience that people who are critical of one are very often critical of the other as well.
Similarly, it's been my experience that people critical of Palintar are also critical of the organizations and governments who use the services of Palintar.
The reason why Palantir is the "tech boogieman" is because it willingly and enthusiastically sells its tech to governments specifically for surveillance reasons, and even consults them as to how best integrate it all. And then we look at the politics of its owner, and it's explicitly anti-democratic - so it's not a coincidence.
And yes, of course, complaining about Palantir online isn't going to change the government policy. But if working for Palantir means that no other self-respecting software engineer will want to shake your hand, then fewer people will want to work there, and their surveillance tech will be lower quality and have more holes in it.
> I think this is position is a bit naive. Like saying "but child porn is just bits, like any other type of file".
Everything has a domain it operates in, so I believe that statement has nuance. In particular when you look at how general or specific the domain is.
From the perspective of something uploaded to S3? It sure is like any other file.
From the perspective of a website dedicated to the dissemination of child porn? It clearly is more than "just bits" in that context.
How about from the perspective of a search engine, where that site may be indexed? Welcome to the grey area that all of these debates are rooted in. Technically it's just indexed strings or ngrams, so it is like any other type of file. But there's an argument that a search engine should "know more" than just the raw data, and should be able to understand that context somehow.
Palantir is in this grey area. The software isn't built for spying. It's built for managing and understanding vast amounts of data. Can this be used for spying? Yes. It can also be used for double blind clinical trials. Or maintaining insurance systems. Or coordinating disaster relief efforts.
It operates technologically in a very general domain, but their flexible user-defined ontology makes it very powerful at operating in more specific domains. So it's a lot less cut and dry than the dissenters make it seem, IMHO.
I’ve only worked with a handful of LEOs in my time there, but the ones I did work with were genuinely concerned about it.
If you’ve never worked with a government agency before, let me tell you, Hanlon’s Razor is in full effect. Never attribute to malice that which is adequately explained by stupidity.
The higher ups know that free reign for their officers to look up anything is a recipe for legal nightmares. The individual officers are often just too dumb to realize that they shouldn’t search for something, and do so out of curiosity. Having blocks in the way like requiring a case number be entered to run a search, and knowing that that search and results are forever associated to that case, cut down a lot on abuse.
Are all forms of abuse the same, or so easily mitigated? No, but I’d like for some system to audit that and at least try to enforce it.
Not sure why you're being downvoted. All your comments have been very well and politely argued (whether people agree with them or not) and this is the first time I got a pick at what palantir is actually doing. Thanks for the info.
Meh, it’s fine, downvotes don’t bother me. Especially when you take the position that goes against common opinion, it’s to be expected.
The irony to me is that when I was at Palantir they talked extremely openly about all their technology. A ton of it was open sourced, and in depth tech talks were posted to YouTube, but almost no one watched them.
I think people just like writing them off as shady or sketchy because it’s easier than needing to reason about the clear ethical nuance that exists in the government contractor space.
During a training, demonstrating how to do geo searches on license plate data from ALPRs, a detective insisted we look up their car.
The hits were all more or less expected, a bunch around their home and the precinct, local supermarkets, etc. But then a weird grouping way out of town. Detective insisted it was a mistake and wanted more data, thinking someone stole his license plate (insert eye roll here).
Anyways, zooming in and looking at the highest density of hits, it ended up being a strip club that he frequented. He got bright red and his buddies didn't stop giving him shit for it all day.
This right here is a great argument for why "helping them use [the data they already have] more effectively" (as you've said) is probably not as great a mission as you're making it out to be.
Well, in the context of this post, that is irrelevant at best, or whataboutism at worst. I notice you don't offer a defense of Palantir here. Why is that?
> Well, in the context of this post, that is irrelevant at best, or whataboutism at worst. I notice you don't offer a defense of Palantir here. Why is that?
HN won't let me reply to that, so I'll do it here. I didn't feel the need to offer a "defense" of Palantir there since I already outlined it above and don't feel like there's value in repeating myself.
But you asked, here you go - The government is going to use the data regardless, so long as they have it. The only effective recourse is to make them collecting or storing the data illegal, but good luck with that, seeing as even when it is illegal they do it anyways (see: Snowden).
Do you think is Palantir weren't there they'd just say "Ah well, I guess that's that" and be done with it? Nope; they'd go to Lockheed, Raytheon, IBM, or another long-time contractor to build a replacement almost immediately. The technology is genuinely nothing special. It federates searches across disparate database, and stitches results together.
But for me, having worked at Palantir, I know what mechanisms are provided to attempt to mitigate abuse. So as long as that data does exist and is going to be used, I'd prefer to have it be as well controlled and audited as possible. And sure, you can argue that the government may forego auditing, or be entirely corrupt, but that doesn't seem to me to be a good reason to not use tools that at least provide that capability.
And we would think the same about Lockheed, Raytheon or IBM. Just because there'll always be someone without morals doesn't mean we just shrug and say anything goes.
I’m not saying we should shrug and say anything goes. I’m saying all these companies share one commonality - the are building things for the US government.
If you want to see change that is more than just superficial, that’s where you need to make it.
I'm talking about neutering the surveillance state. That's not something that demonizing individual companies is ever going to do. You need to push for that change at the government level.
Unless you're pro-surveillance, but your previous comments made it seem like the opposite.
Sure. That doesn't mean that what Palantir are doing isn't wrong - selling tools to an entity which is going to use them for harm is wrong, even if they might result in slightly less harm than selling them some other tool.
If someone comes into a gun shop and tells you they want to shoot up a classroom, selling them a pistol instead of a machine gun doesn't make you immune from judgement.
What makes Palantir different than any other software vendor that sells to the government?
Want to know how a lot of that data is generated? Microsoft excel. Arguably without excel there would be less data to act abusively with. If only Microsoft just refused to sell excel to the government.
People are happy to accept that excel is general enough that it isn’t made for that one purpose, but refuse to apply the same reasoning to Palantir; that may seem like whataboutism, but they genuinely are comparable tools if you take away the marketing and fear mongering.
As for your analogy, it’s more like you own shop that makes metalworking tools, and someone buys some. You don’t know what they are going to do with it, and (here’s where it gets opinionated) you shouldn’t need to care. You sold a tool to someone. Can the tool be used to make a gun? Sure. But it can also be used for anything else involving metal work. At some point we need to accept that the responsibility for how a tool of used needs to be placed on the person using it.
Palantir develops tools to be better at tasks that are pitched to them in some degree of detail through the RFP process and similar processes. Palantir as a company would not exist if it were not for the less savoury of those tasks.
We do not hear, "Government X purchases a couple thousand Excel licenses to keep a database on people it'd like to kill", and I imagine that Microsoft does not sell Excel to Government X in response to an RFP for tooling to keep track of people it'd like to kill.
(Of course, it'd probably stick its metaphorical fingers in its ears if it did hear about Excel being used for that purpose.)
> The only effective recourse is to make them collecting or storing the data illegal, but good luck with that, seeing as even when it is illegal they do it anyways (see: Snowden).
So, basically "screw rule of law, people are gonna do it anyway, so let's make money off it!"
Do you buy things from amazon? They treat their warehouse employees terribly, so I hope not, lest you facilitate an immoral act.
Do you ever use uber? They work to keep their drivers legally declared as contractors, despite treating them employees in all other regards. That’s pretty immoral, so I hope you don’t encourage it by supplying them with business.
Machine Learning/AI adds an entirely new perspective on big data. It can see patterns in data that would be meaningless without it, correlations that can't be found in other ways. Some data is OK for a government to have if they can just look it up in exceptional cases, but not if they act on every single bit of data collected.
In essence, using ML is creating new data (or at least: Information from the raw data). This information can be very revealing and not at all in line with the scope the data was initially collected under, which is one of the principles of GDPR.
You can't really view data as static once it's collected. Combining with other existing data and new methods of data analysis make it a totally different picture in terms of privacy.
I interned in St. Louis one summer and my neighbor worked at Boeing. The only reason I remember him to this day is because he introduced himself as a programmer working on smart bombs. He immediately said it was okay because he didn't fire the bombs, he just made them smarter so they didn't kill innocent people.
Is it a realistic expectation that we have no one working on smart bombs/weapons technology in general?
I’d love to be in the utopia world where there is no war and the entire planet gets along. But even the US has political bipolar disorder, so we’re a looooong way from global peace.
I do think we spend an absurd and inappropriate amount on defense, but as long as we are forced to develop weapons due to the global geopolitical climate, I’d consider it preferable to work on making them more targeted.
My problem is I don’t like or agree with the people picking the targets, but that’s a whole other argument.
Maybe you can’t convince everyone else to behave ethically, but you can hold yourself to that standard. Working on bomb tech is over a line for me personally.
Sure, I get that. And I think it’s a position shared by a lot of people. Probably the majority of people. But it’s an ugly necessity of modern politics.
We could absolutely stop all weapons production in the US, I’d just ask for some time to learn Mandarin first.
As someone involved with the industry, this is my take.
The work defence contractors do and will continue to do is a measurable positive in the world. This work generally falls into three categories:
First is projects that make the lives of armed forces personnel easier. There is no moral or ethical hazard with these projects. They simply are making it so that pre-existing work, often non-combat work, is less tedious.
Second is projects that actively protect the lives of armed forces personnel and assets. Why would somebody have any reason to feel anything but pride in their work when they know that the only things it could ever do is save lives.
Third is the arguably morally shaky stuff. Weapons technology would fall under this category. The thing people seem to miss about this category however is what new projects' goals are. Very rarely is it ever to create a more lethal weapon. Almost all development of this kind focuses on either making existing technology cheaper, more precise, or have increased range.
In the first case, you could argue that it makes the weapons more likely to be used which is fair however if a weapon system is in active use already, it is unlikely that a cheaper version would do anything other than bring down operating costs. In the second case, these projects are actively making these weapons safer. Look at the Hellfire R9X as a prime example. This weapon sole purpose is to minimise casualties and minimise damage to the surrounding area. Finally the third and last case I mentioned. Increasing range is important for force projection reasons but the most immediate benefit is that it allows armed forces personnel to be further from danger. These people would be in this fight regardless but now they are able to operate from a safer distance instead.
Weapons tech gets a bad rap for obvious reasons but I see that as mostly having been a relic of a past era. Nowadays there isn't really any reason to make more lethal weapons. We already have those and you can see a steady trend in new military technology towards lower risks, costs, damage to the surroundings, and minimising civilian casualties.
Additionally, in my experience, you generally have a choice whether you work on a project or not. Leadership fully understands what they do as an organisation and get that some people may not be comfortable with doing certain projects. If you can't conscientiously work on a certain type of project, simply say so and leadership will take that into account when planning personnel assignments for projects.
TL;DR Armed conflict and existing weapons will never go away. What we as a society(namely the modern purpose of defence contractors) can do instead is make armed conflict less dangerous to those caught in the crossfire and those who risk life and limb to defend our countries.
Even those first two can have fairly foreseeable negative effects, though. Not necessarily net negative, mind you, but negative. If (e.g.) American soldiers have less tedium in their downtime (e.g., because paperwork is automated) and are safer (e.g. because their body armor is better), it could very reasonably lead to missions or even whole wars being undertaken that would otherwise be unconscionable. It's kind of like boxing gloves or football pads - on the face of it, it makes things safer, but because everyone knows it's "safer", they're willing to push it further.
Not judging your decision (although I personally prefer not working for defense/defense contractors), just pointing out that "it only makes soldiers safer!" isn't an ironclad rebuttal.
That's the point of armies, to be as lethal and effective as possible. A protracted war benefits no one, and a powerful army acts as a deterrent. I think that if you think your country has the right to self-defense, you have a moral obligation to make it as lethal as possible. Ethics is not an excuse for inaction or abandonment of responsibility.
I think this calculus becomes slightly different when you're talking about a nation that will never be in an existential defensive war. A protracted war benefits no one, but knowing that a war would be protracted might benefit the side that would otherwise be the loser by serving as a deterrent to the winner.
A protracted war benefits the military of the weaker side but I doubt their people would like it. Also, I can't recall the last time the United States was deterred from entering war because it would last too long.
If one considers the whole operation of armed forces to be unethical, it's easy to find ethical and moral hazards with the first two points as well.
Analogy: Would you consider it unethical to provide tools and services that make the lives of human traffickers easier? To actively protect assets of drug lords distributing contaminated opiates and meth?
Slave- and drug trade are about as likely to go away as armed conflict.
I think this argument is a bit unfair. As much as I would love for our armed forces and the concept of armed forces as a whole to be unnecessary, the reality is that there is no realistic way for nation states to coexists without some type of armed forces as a deterrent to conflict even if they were never to be put into action. Because of that I don't think it is possible for the simple existence and operation of armed forces to be unethical. Sure armed forces can be used unethically and individuals or units can do unethical things but that is true for all organisations.
This is in the same way that corrupt charity organisations can defraud the causes they claim to support and corporations can knowingly risk the lives of their workers due to either negligence, mismanagement, or for the sole purpose of saving costs or making more money. This doesn't make charities or corporations unethical solely because they have the potential to be.
Additionally, armed forces can and do maintain their operational fitness by using their immense logistics networks and pools of human labour to do very real quantifiable good in the world. Outside of their use as a deterrent and show of power, the US armed forces spends a large chunk of their time and effort responding to natural disasters, taking part in humanitarian efforts, and building infrastructure.
---
On to the analogy, I would like to reframe both examples a bit.
For the purpose of this discussion, illegal transportation of individuals falls into two categories: Human trafficking and human smuggling. Human trafficking is done without consent of the individuals being transported. Conversely, human smuggling is done with consent. Despite being illegal, human smuggling provides a very important service to individuals. It may be against the law but it can be argued that it can provide a net benefit.
Mind you that it comes with severe downsides and by no means am I trying to minimise those. I'm just trying to keep this from turning into a 20 page essay on the topic. One of those downsides namely is that it increases the volume of people being transported which makes it cheaper and easier for human trafficking rings to operate. Another downside is that a large number of people die during transportation, both those consenting and non-consenting.
Now if you move onto the less clear cut outcomes, human smuggling allows individuals to enter a country without going through the vetting process. This opens a vector for criminals and terrorists to enter a country however at the same time, it provides a means for people who either aren't willing or aren't able to wait untold years to get a visa to enter the country only for it to then not be renewed a few months or years down the road. These may be individuals or families that want to be productive members of society in a country with a higher standard of living or are trying to move to a location where they can be treated for medical issues that they can't be treated for at home. Another common case in human smuggling is individuals using human smuggling to cross countries that have closed their borders to them so that they can get to a country that will accept them as refugees. The number of reasons somebody would illegally enter a country go on quite a bit and the list is full of both morally sound and morally reprehensible reasons.
With that in mind, while I personally wouldn't participate in any type of business connected to the illegal movement of people across national borders, I see no moral or ethical issues with someone who does so for the sole purpose of making the process safer. Despite the fact that it is illegal, the way I see it is that you are reducing the risk that people lose life and limb.
---
Moving on to the other example. My opinion is that most drugs should be legalised and regulated. By moving to a legitimate market and starving these organisations' source of income, we should see drug lords be replaced by corporations. You could argue that the cartels would jus...
I generally agree with your comment but this part is IMO delusional:
> These people would be in this fight regardless but now they are able to operate from a safer distance instead.
I sincerely strongly doubt US would be fighting 5 wars with boots-on-the-ground in Afganistan, Pakistan, Somalia, Yemen and Iraq, if it weren't for remotely-operated drones.
I don't believe it is unethical to work on weapons technology any more than I believe it is unethical to be member of the armed forces. We live in a world with bad actors, so some amount of work on defense is necessary for survival.
However, I also believe you are responsible for the consequences of the technology you create. If you decide to work on smart bombs and because of a bad government those end up killing innocent civilians, some of that blood is on your hands. If you're willing to accept the risk of that, I respect your choice.
Personally, I would rather avoid that by not working on that category of technology at all. It's not the kind of mark I want to leave on the world. I do appreciate that I only have the luxury of this choice because I live in a country where others do choose to work on defense so that I can be protected.
> However, I also believe you are responsible for the consequences of the technology you create.
I agree with this to an extent, but at the same time, technologies you develop can go well beyond what you intended them to be. Do you think Ritchie knew C would be used the way it is today? It is used in UAVs, smart missiles, and more. Building faster algorithms also means faster weapons. Everyone working at SpaceX is directly or indirectly working on missile technology. You can do this with practically any technology, even as simple as building a better screw. Building materials that are stronger, lighter, and cheaper reduce the prices of homes, but it also has applications in war. Studying diseases and vaccines directly impacts biological warfare research.
It is easy to say you're responsible when you're working on things like smart bombs. But it isn't easy if you're researching fertilizer. The thing that arguably has saved the most lives yet how many lives have bombs and bullets taken?
> Personally, I would rather avoid that by not working on that category of technology at all.
So what I'm saying is you're working on that tech, just how removed are you? I also don't think there's a cognitive dissonance. You can be pro putting nitrogen in soil and anti explosive nitrates. But saying the two aren't related is naive.
Engaging in any kind of commerce (i.e. beyond growing food for yourself and your family) means paying taxes, which means funding governments and armies.
If you're working on weapon systems that will be used to attack, say, Nazi Germany, and you know or expect that they will also kill innocent civilians, then you are de-facto endorsing that murdering[0] innocent civilians is worth it to stop Nazi Germany, which is a resonable, if debatable, position.
If you're working on weapon systems that (you know or expect) will be just be directly used for murdering[1] innocent civilians, then you're endorsing murdering innocent civilians.
If you don't know where on that scale things are, then you're reasoning under uncertainty, and your ethics are going to have to deal with that.
0: technically most of those would actually be manslaughter, going by the usual definitions, but a: not all, b: that's not a verb.
The "if" is significant. If you work on weapons that turn out to be used to stop legitimately bad actors with minimal collateral damage, then your hands are as clean as anyone's can be in war.
I don't generally believe in black and white, so I'm making no claims that anything is 100% ethical, 100% unethical, 100% blood on your hands, etc. I think it's possible for good people to work on weapons and still sleep soundly at night. I also think it's possible for good people to work on weapons and end up regretting the consequences of that choice.
I mean, that seems legitimate to me. Various governments are going to continue dropping bombs. If there's a market for PGMs for U.S. operations, that means less collateral damage; whether you like to think of it that way or not.
It's rather amazing the mental gymnastics people perform to get out of cognitive dissonance. I'm much more sympathetic with people who know exactly what they're working on and the full effects of that work, why some will find it distasteful, but nonetheless can at least make a case for why those people are wrong and the work important, rather than agreeing with them but hastily adding some transparent excuse to avoid being lumped in the distasteful category.
Yeah, and the commenter upthread didn't personally saw Jamal Kashoggi up alive, and he's got a whole headful of justifications as to why he couldn't possibly be complicit in that. When he is. He doesn't have a single atom in his body with the ethics and morals of Tim Bray...
I could argue the exact same about literally any database used by the government. Their tools are just easier to use and more expressive.
If you don’t like the bullshit the government does, that seems like an issue you should take up with the government, no? You’re using a tech company as a scapegoat.
> Their tools are just easier to use and more expressive.
And that's the problem – they're giving governments tools that let them more easily and effectively infringe on the privacy rights of their citizens. Privacy isn't invaded when the data is collected, although that's a necessary step; it's invaded when a cop or bureaucrat queries that data to see what a citizen has done. Giving them that tool is morally complicit.
~ "That's not my department," said Wernher von Braun. ~
This is a ridiculous stance to take. If true, then open sourcing any form of software is morally corrupt since the government could then use it for evil.
Quite frankly, I don't think any reasonable person would consider selling software to the government to be morally corrupt unless they were reasonably confident the government would use it for evil. Right now, nobody has made a compelling-enough argument to make me believe the government will use this software for evil.
Intent matters. Palantir specifically makes surveillance tech, and specifically markets it to governments.
As for compelling arguments... is the history of most of the world's governments not enough for you? In US, you can look at the census for a case in point: this data was used to chase draft dodgers during WW1, and to compile list of Japanese for internment during WW2. Curiously, by WW2, the federal census law had specific provisions preventing the Bureau from disclosing that information to any other government agencies, precisely to prevent this kind of use - Congress simply repealed those provisions. Nevertheless, the Bureau subsequently denied sharing that data and buried any leads, so we didn't have definitive proof until this century.
> This is a ridiculous stance to take. If true, then open sourcing any form of software is morally corrupt since the government could then use it for evil.
As soon as the data is collected it is going to be used. That is where the privacy invasion is. If a massive data set on citizens leaks and becomes available online, people will still consider it an invasion of privacy, even if it was never "used" prior.
The same argument is actually used w.r.t. DHS/ICE using GitHub to work on software to better track illegal immigrants.
This is software that will ostensibly cut down on errors and speed up processing times for things like asylum claims, but for some reason Github was called out as being complicit.
I think the argument against Palantir is better formed than an argument against Github based on specificity of the tool, but I think both are driven more by viral outrage than careful consideration of any moral calculus or actual hands-on knowledge of how the tools are used.
Palantir is _explicitly_ contracted by a government to create a tool for combining surveillance information to make tracking individuals more effective.
Github is providing tools (Github Enterprise IIRC) which is presumable being used to create tools which are used to track and deport individuals.
Some databases used by governments are for things like making sure people driving cars have received the required training/certifications to share public roads, or that their vehicles carry the necessary insurance.
Other databases get used to track down, torture, and murder critical journalists - by dismembering them with a bone saw while they're still alive.
Remind me again which kind of database Palantir staff work on?
Well, when I was at Palantir I worked with police departments integrate a dozen or so databases so looking someone up took one query rather than several, cutting down on the time needed for traffic stops drastically.
The software was also used to run double blind clinical drug trials - the acl granularity was great at that; you could have different roles for drug manufacturers, doctors, and patients, none of which had absolute information. Then trial supervisors could open up the data for analysis at the end of the trial.
At that same time, the ability to use GPS phones to upload data over unreliable networks was used when they teamed up with the Clinton foundation and Team Rubicon to improve disaster relief coordination for hurricanes.
Myself or coworkers (“Palantir staff”) worked on all of these.
It isn't simply a database, as I'm sure you know. There are multiple pipelines of information which need to be glued together, cross correlated, etc. They (Palantir) provide services to make policing citizens easier and allow police departments to overstep their boundaries.
Also, I'd like to point out that this isn't an "either or" situation. Obviously the government is involved, as they no doubt put out the contracts, but Palantir is the one cashing the check. They've created their entire business model around big secretive government contracts.
I would treat this app as advanced kleptographic malware after knowing that they're developing this app with the NHS to contact trace millions of people to 'stop the spread' on a centralised server.
Can someone please explain to me how a “libertarian” like Peter Thiel could start and run one of the most anti libertarian companies? What are his beliefs, exactly?
I mean besides “zero to one” and “competition is for losers, build a monopoly”?
I wonder how much effect he had on Zuck early on. He was the first big check in. Seems it was a great match there.
Principles dont matter if you dont have the money to put them into practice.
If palantir is what it takes for him to fund/back a visionary company or leader that will create a truly freer society, then that investment was well worth it.
Because his wealth isn't enough and he can still insert himself into other policies to obtain more influence, more power, and more money?
My question to you is why you care what Thiel thinks and why you're dissapointed. As far as I'm concerned, these guys are the lowest of the low and they only hold their position through gross exploitation.
Nobody should be amassing this impossible wealth and seeking to gain more and these CEOs are basically economic despots to my mind.
There's libertarians, and then there's libertarians. Thiel is a huge fan of Hans Hermann-Hoppe, who writes things like:
"In a covenant concluded among proprietor and community tenants for the purpose of protecting their private property, no such thing as a right to free (unlimited) speech exists, not even to unlimited speech on one's own tenant-property. One may say innumerable things and promote almost any idea under the sun, but naturally no one is permitted to advocate ideas contrary to the very purpose of the covenant of preserving and protecting private property, such as democracy and communism. There can be no tolerance toward democrats and communists in a libertarian social order. They will have to be physically separated and expelled from society. Likewise, in a covenant founded for the purpose of protecting family and kin, there can be no tolerance toward those habitually promoting lifestyles incompatible with this goal. They – the advocates of alternative, non-family and kin-centered lifestyles such as, for instance, individual hedonism, parasitism, nature-environment worship, homosexuality, or communism – will have to be physically removed from society, too, if one is to maintain a libertarian order."
This quote is from Hoppe's "Democracy: The God That Failed", which Thiel has referenced before. The overall thesis there is that economic freedom is the cornerstone of libertarianism, and it's fundamentally incompatible with democracy, so libertarians have to explore other options. It specifically promotes monarchy, on the basis that a monarch is more like a business owner with a vested long-term interest.
You might argue that this isn't really libertarianism - indeed, many libertarians don't consider Hoppe to be one. This is basically proto-neo-reaction, at the point where it still hasn't explicitly rejected libertarianism altogether. But regardless of where you draw the line, this all is the evolution of Murray Rothbard's strain of libertarianism, after it moved away from social liberalism towards paleoconservatism - it's not just some random people coming and declaring their politics to be "libertarian" out of the blue. Nor is there any shortage of ex-libertarians in NRx circles, so it's clearly a common path, not unique to Thiel.
Needless to say, the libertarian paradise described above could use something like Palantir to implement the "covenant".
Sounds more like fascism with the alignment of political and corporate power.
But the label is irrelevant, Thiel and Palantir's actions are anti-democratic and their involvement in any government activities immediately raise privacy and safety concerns.
Just because Thiel has referenced this speaker before doesn’t mean he agrees with everything he says. It’s a bit suspicious Thiel would agree that talk of “promoting“ homosexuality should get someone expelled from society.
Do you have any references showing Thiel believes these words you quoted?
"I no longer believe that freedom and democracy are compatible"
This is literally the primary thesis of Hoppe's book.
And note that Hoppe's isn't saying that promoting homosexuality should get one expelled from society, but rather from communities with "covenants" that restrict it. This arrangement is all about one-dollar-one-vote, so rich people needn't worry - they can live however they want, even establishing their own "covenant" if need be.
That explanation is much more reasonable. If people can choose which covenant to belong to, that's better then some larger collective (like a federal government) imposing it's morals on smaller groups.
>Can someone please explain to me how a “libertarian” like Peter Thiel could start and run one of the most anti libertarian companies? What are his beliefs, exactly?
None. He is a slimy character with no principles or morals besides his own selfish drive. He wants to make money, period. Other considerations don't factor into it. Make no mistake, he (and people like him, the capitalist sociopaths that hold power in our society) would kill your entire family for a handful of pennies if they would get away with it.
If there were any justice in the world these kinds of people would be shunned and censored by society. Yet it's precisely the opposite. These are the behaviours which are encouraged and rewarded.
Contact tracing isn't bullshit. It is how they dealt with sexually transmitted diseases four decades ago...it works.
The issue is that the UK govt is composed of a combination of incompetent and corrupt people (one of the companies involved in this is owned by a minister). And they are trying to go with the low-cost option (whilst the govt is subsidising 1/4 of salaries paid in the whole country) rather than the old-school, proven way.
It is kind of shocking but it is the result of a political system (and culture) that rewards incompetence and sloth.
The Australian app is reportedly based on an open source app from Singapore. The Australian government said it would release the source for its own version in a couple of weeks. The code is GPL, if it's this one: https://github.com/OpenTrace-community.
Can anyone point to some articles of how Palantir is abusing data?
The reason I ask is, I also thought that it's "just supreme evil" until I listened to a podcast with Thiel (Dave Rubin podcast) recently... In it, he said that he started Palantir as an experiment, what can be done against organised crime/terrorism while respecting civil liberties and privacy. His argument was, that we do need to stop terrorism, because people (not HN crowd, but the public at large) are obviously not willing to compromise safety for privacy, so each terrorist attack results in privacy- & civil liberties-curtailing laws - so the question is, can we use data/ML/etc. to safeguard our society while preserving as much privacy & liberty as we can.
But I've almost no real (and insider) info about Palantir to judge whether that's actually what the organisation is, 10 years or so later.
Having said that I know nothing of Palantir, today was the first time I heard of it.
But I would never ever believe anything a CEO says in the media that can't be validated. Their very job is setting the company in a positive light above all else. Especially when it comes to vague terms about company ideals that can't be proven. And 10 years is an eternity in IT. Remember Google had their motto "Don't be evil". Look how that turned out.
This is why I mentioned the time frame - this goes into trial EDIT:TOMORROW* (still don't think that gives much time to audit the source code) in Isle of Wight, and ITV news said it is going out to the rest of the UK Thursday - so its being installed without any code being shown yet.
You don't think the 5 eyes don't get handed out mobile data location directly from the cariers? I am more concerned about a database that will inevitably be accessed by lots of researchers, and of course one will think it's a good idea to store a copy on a USB key or a public s3 bucket, etc. It the same story happening again and again and again...
This is a legal framework, that is what is happening now. I noted this a while back in a flippant comment "Capitalism with Chinese Characteristics" (which didn't go down well :).
A contact tracing system and a virus with n days of non-symptomatic incubation, and international borders. Let's think this through a bit. Of what benefit is to the traveler admitting/host country (H) to note on the "medical passport" issued by origin country (O)? Obviously, for this thing to even work on paper, countries need to exchange massive amount of information.
The fact stands [challenge it!] that the socio-political regime being rolled out in guise of defense against Covid-19 and future viral friends aligns perfectly with a specific ideological position. It really was 'Nature's gift'.
What used to be accomplished via violent revolution is being accomplished with 'scientific' efficiency and very little blood. Oh, that figurative blood in the streets of middle and lower classes? "The Virus does not discriminate!".
But imagine the new frontiers it opens! Countries would exchange and trade massive amounts of user data like oil. Want to participate in air travel? Better to sign treaties to buy tracing data, as without it your country won't be compliant with the international safety laws. I see trillions of new wealth being built there. On a serious note, this reminds me of a quote from one book: "the fourth round will be remarked by a fierce battle between materialistic and spiritual forces in humanity". We're like being subjected by a high voltage electrolysis now: those with moral are being dragged to the right and those without it - to the left.
I just got an urge to resurrect Frank Herbert to see what he has to say about all this. Wonder who is our dear Leto the God Emperor, the benevolent monster who is dragging an un-surprisingly quiet and docile humanity to its designated future. One hears the monster has 7 heads.
I'm not sure about the hand me down binary model conclusion. We're in agreement regarding the inevitability of this turn of events, but then again it was apparently clear as day thousands of years ago to "self isolating" monks.
But the clarity of the choice is stark, and like a blast of cold air invigorates one's morals and character.
In a way I think this will actually hamper their efforts.
After all: With the corona tracing apps, people are becoming much more aware that their data is used to trace them. A lot of people I know are talking about leaving their phones at home when they go out.
Considering the carriers have had this data forever and are probably sharing it, this awareness is a good thing for privacy.
Stop spreading FUD. Governments have had access to your location and associates for decades via the mobile phone network.
I've worked at a telco where we were analysing customer's behaviours and almost everyone follows a pattern at some point e.g. going from home, via public transport to work. The resolution isn't high but over time you build up enough samples to accurately determine exactly where they live, work etc.
Yeah, but that is a small part of contact tracing (with law changes needed in some countries). The next step is to have personnel to notify suspected contacts, tell them to isolate themselves and then test them.
You need to have enough people for that, you need to have the testing capacity and you need to have people following that order.
It will work well if you have very low community spread and can start tracing on cluster level. If you are not there it can still be a good tool but you would probably still need a lot of other restrictions as well.
Collecting vast swaths of data to know where I live and work, which are literally the first two questions on my income tax return btw, is very different from knowing the name of every person I've been near and the location where I'm near them. We've all seen how these "geofence warrants" have gone -- a lot of innocent people being dragged through the court system for absolutely no reason.
If we do end up implementing a centralized contact tracing system, the results will be very interesting. I would love to know the name and location of every lobbyist that my elected representatives have met with, and I'm sure that in the name of public health they won't make any efforts to thwart this tracking.
This is exactly correct. I helped create the system used by US telcos to track users. We had GPS precision tracking on all users and used big data to geolocate that data against other users, retail establishments, and road locations. This was not nefarious. We used it to find and fix tower, highway issues and market to companies. Users agreed to it as part of their service contract.
We were required to sell the data to the government on demand due to CALEA and similar legislation.
Those who complain endlessly about AANG largely seem to have no idea of how our communications networks actually work.
Your location, activities, banking, etc. are not private, will never be private, have never been private, and it has nothing to do with AANG. The 3 AANG that I have worked are woolly lambs. That F guy is a bit suspicious, but I don't have first hand knowledge.
For a start, installation of the app is voluntary so if you're planning something you really don't want the government to discover, you wouldn't install it.
Also, intelligence services already get call metadata so they know which mobile towers you are using and probably lots else.
As for effectiveness, I think it will be useful. There have been several papers crunching the numbers and it looks like it will have an effect. It doesn't need to be perfect - any non-trivial reduction in R is helpful.
They mentioned a few weeks back that the end goal would be if you had the app it could act as a virtual health passport and you’d be allowed to go out shopping, go to work and the pub, etc. And without it? Stay at home. So yeah, technically voluntary, but, y‘know...
Install it on a secondary phone? Of course it all depends on how serious the government is and how hard you are willing to work to circumvent controls.
Just a minor annoyance except of course for those not rich enough to own two phones, but I suppose if you didn't want to be oppressed you should have thought of that before you decided to be born poor, right?
"The Secretary of State has also committed to making the source code open source. That may not happen immediately on release because the fabulous NHSX development team are all working hard on getting the product ready. But it will happen."
Sounds like there is an intention to deliver on the source. Bigger question is whether the server side is open sourced (as clearly that has limited value for verifying what is happening remotely).
> That may not happen immediately on release because the fabulous NHSX development team are all working hard on getting the product ready. But it will happen.
That's just annoying nonsense though isn't it?
There are reasonable (well 'ok they happen, whatever') causes for delay in open-sourcing something that didn't start out that way, but... Do any of them take more than seconds of developer time? They mean it 'may not happen immediately, because it's stuck in legal', surely?
Sounds like there is no intention to deliver the source on release, and once the release has been out for a couple of weeks, the benefit of public review of the source code drastically diminishes.
The source code should be released before the public release of the binaries, not after, and not doing so means they're trying to hide what the app does from timely public discourse.
Releasing a tarball (even if it lacks the tools to build it) isn't that hard, is it?
Sure. It's maybe a bit unsettling to see a global corporate monopoly telling countries what they're allowed to do, but in this case it's the right thing.
I trust a democratically elected government above a duopoly personally, and I'm not convinced the decision of how far to trade off privacy concerns against public health is one those corporations are competent to make.
Google and Apple need to advocate privacy for reasons. But they also want to help the government (PRISM example). So they carefully craft a bug that only government knows about and that allows the extraction of location data. If ever discovered it was not intentional. The NSA backdoor at RSA is an example [1].
I think it’s unsettling that were so used to companies and people with actual power doing fuck all to stand up to government abuses that it’s weird when it actually happens.
Will be interesting to see if they can ban Indian govt. app[1] which needs full location access(clarified)[3]. A lot of people like this app(including me) but also know government does not have good track record in securing private data.
Previously Apple were made to bend their rules when India threatened to ban Apple devices if they don't allow TRAI Do Not Disturb app in 2018.[2]
Thanks, just looked it up as well. But to the GP’s point, I do hope Google/Apple start banning apps (especially released by governments) even if they don’t use the contact tracing API.
Mostly because it was released very early. And for a country like India, it may prove advantageous in coming months. We do not have any vaccine for Covid19 and can't be in lock-down forever. We have to learn to live with the virus for few months at least. Apps like these can help in contact tracing while allowing many to live normally.
Also I personally think the permissions(location and bluetooth) are fine for an app like this to really function. I have read someone mentioned on HN that these platforms prove their worth when >60% people are using them(I maybe remembering wrong though).
Not sure about Android, but if you're on iOS you have the choice to disable "full location access". In case it's needed to run certain feature in the app, you can choose the "Allow while using" option.
I do that will all the apps that require location access: local food delivery, cab services, vehicle rental and what not.
We can and for most apps I do this too. But some apps, like this one, require the location or they won't even start. Also at least on android(which also has fine-grained permissions now), this app specifically requires on-going(background?) location access.
Recently this has also been made mandatory for employees, public and private. So organizations have to ensure all employees have this app on their smartphones. We will see how much this is enforced.
My Android device allows for completely disabling location access (regardless of whether the app requires it) or for allowing access while the app is in use, as you mentioned. I'm never sure if these features are OEM additions or features in standard/AOSP Android versions, so I guess if you're using an Android device YMMV.
Well not if once you use the new API specifically made for contact tracing, right? Since the bluetooth scanning will be done by a lower level system run by play services, the app itself doesn't need the permissions.
As another commenter pointed out, if the app is not using the contact tracing framework developed by Google and Apple, then the app can basically do whatever it wants (mandate continuous location access etc.).
I don't think apps are allowed to do this on Android any more. You can designate that an app can only have location access while the app is open and in use.
Right, for sure, that’s a choice users have. (On the iphone too, fwiw).
Sadly, outside of tech circles very few people would have the knowledge or even motivation to do this. The attitudes of the populace to online privacy are frighteningly callous.
I've been astonished at the number of apps that are grabbing my location in the background. Espescially since I previously considered myself to be quite on top of the permissions I'd granted my apps.
Android for work can reduce privacy leaks; get Island or another app to set up a work profile on your Android phone, and sign into a secondary account on it so it doesn't have your contacts. Now go to settings and disable location access for work apps. I've stopped paying attention to whether or not apps request location data or contacts access for to this. To prevent the apps from running in the background, just turn off work profile when not using it (or look up Greenify's deep hibernation).
Being on a stock, unrooted phone, the thing I miss most is xprivacy's prompts whenever an app wants to use a permission. I just make sure to check my permissions list every month or so to make sure Google hasn't silently allowed an app update to enable new permissions.
and it may prevent the app from functioning properly but the point is that Apple/Google will not allow you to publish an app on their stores which utilises both of location services and contact tracing APIs
Apple logs location history? Other than the wifi location log from ios4 and the significant locations feature (opt in, device only), I don't know how apple is logging locations.
Not only it does but I even remember a story of a crime that was solved in Germany a few years ago just by requesting from Apple the history of the locations.
"I remember a story that is the entire opposite of what you said".
If you have an argument to make based on a specific event, you're going to have to do better and provide some sources. "I remember a story", are we going to throw the full privacy history of Apple away based on your "history" that you remember? Please provide sources or do some research before you comment.
On your iPhone you can go to Settings -> Privacy -> Location Services -> System Services -> Significant Locations and you will see that Apple does keep track of at least some of your location history and tries to analyze it if you don't actively turn it off.
Not true. The iOS location history is encrypted and stored in your iCloud account with a key that never leaves your device. So it's quite different from Google's Location History: Apple stores it, but they cannot read it.
And if you disable iCloud, the data remains exclusively on your device.
That iCloud key only prevents man in the middle attacks. You can still use the key to decrypt the data at rest in the cloud servers, and Apple stores both.
On my iPhone running iOS 13, I had to go a level deeper to find that: Settings -> Privacy -> Location Services -> System Services -> Significant Locations.
To view Significant Locations, I have to provide my Touch ID again. Also, the fine print says "Significant Locations are end-to-end encrypted and cannot be read by Apple." Text repeated in "Location Services & Privacy"[0].
Claims aren't worthless, they'd be breaking the law if they lied. Apple has a pretty decent incentive to not break the law (PR mostly, but also fines), so I would bet they're telling the truth here.
If security researchers couldn't audit closed source software they would be pretty rubbish at their jobs. It's a lot easier to audit open source of course, but researchers do reverse engineer closed source stuff all the time.
Well you can always look at the network packets, most easily accomplished from jailbroken devices. There are many people who look at such matters. It's far from worthless.
Can't Apple push silent code updates to specific phones? That would mean they can access that data whenever they choose to. Even if they don't do it right now, they retain that ability.
Are you talking about the one that won't let you install apps on your own phone without telling it who you are? They're the same. If anything, Apple is worse.
And we'll see how this plays out, perhaps it will be tragically ironic. The majority of people don't ever think about this and are happy to give their location info to Google constantly all day long. This protocol is in almost all ways much less intrusive and more protective of privacy, and the upside for society is potentially high.
It should be a complete no-brainer that everyone who is ok using google or apple maps, i.e. probably 95+% of the country, be ok downloading their local health department's contact tracing app if it uses these APIs. But we'll see how it plays out, in the end it will surely come down to how this is politicized and not actually be based on the technical merits of this protocol at all.
any time I've used garmin over the years it had been only slightly better then useless.
there are other options though. ive been using openstreetmap for about 4 years and have only found the need to use Google once in all that time. although with openstreetmap being crowd sources it really depends on how much work has been done in a particular area but in my country its pretty decent
I have four arguments as to why google are behaving this way with regards to this data. I’ve no idea which, if any, they operate by:
1. It’s actually crap and they don’t want their advertisers to know that their ads are sold based on crap quality data/don’t want the backlash of doing a poor job helping
2. Most people at google care a lot about privacy (either in the secret kept between you and google sense or the more common sense definition people seem to use on hn) and they don’t really think about this data as something the firm has/should release
3. They are afraid that if governments realised they had this data then google would be regulated or every minor security agency/random government department would be demanding access to it by law.
4. They strongly feel that surveillance by (well intentioned?) private companies is ok but by governments it os not.
> 2. Most people at google care a lot about privacy (either in the secret kept between you and google sense or the more common sense definition people seem to use on hn)
When I worked at Google, I saw a lot of earnest efforts to keep data private, in a way that really was sufficient (k-anonymized with large k, for example), but in ways that have no outward proof that it was happening. Send all possible data to the server, then make sure it's properly clustered and scrubbed before it gets stored or analyzed. And it's not easy to explain to someone who can see the whole system that from an end-user's view this is identical to just scooping up everything.
I can verify that, from an outsider's perspective, this is indistinguishable from just scooping up everything and saying "trust us, we care about your privacy".
I expect google buys data from data brokers and this is counted as sufficiently anonymous to not be included in the location history google shows. But this misses the point because looking at your location history shows you the quality of a tiny fraction of Google’s location data and don’t tell you much about the rest of the data
I work at Google but don't have any visibility into this project, nor do I speak for Google on this. My uninformed guess is that the answer is more precisely:
5. They very much want this project to succeed so that it can save lives. They know privacy concerns are (rightly) an existential threat to the success of the project, so they are trying to address those by drawing this firm line.
It's not 1 or 3 because anyone can look at the location data Google gathers on them and judge for themselves if it's crap. Just search for "view google location history" and it will take you right there.
I think 2 is pretty close to the mark. Despite HN's understandably cynical take, everyone I talk to at Google cares a lot about user privacy.
Why do people/google suddenly start caring about privacy now?
You can only see the tiny fraction of data that relates to you, and only the data that has been identified to you. This says very little about the quality of the rest of the data google have. Doing the double slit experiment once with a single election will not tell you much about quantum mechanics.
We probably shouldn't call them "contact tracing" apps since what they plan to do is so different than manual contact tracing. "Exposure notification" is a better term.
Nothing prevents anyone from using their phone's location history to remember what to tell the contact tracing people.
Contact tracing: Alice is able to say, “I was in contact with Bill and Carol.” Then authorities can talk to Bill and Carol, and have them trigger their phones to see who they’ve been near. But because that’s slow, most plans would upload the lists of who’s been near who to a central server. Then the authorities can do a simple query to see who’s been near who.
Exposure notification:
Alice enters a code that she got with her positive test result in to the app. The app has been continually broadcasting rotating, random identifiers which it then uploads to the central service. The code she entered verifies to the central service that she has a legitimate positive result. Bob and Carol’s phones periodically check with the central server for the list of positive IDs. Their phones stored one of the IDs from Alice’s phone when they were near each other earlier. Once they get the latest list of infected IDs, their phones will alert them that they have been exposed and should be tested.
In CT, the central service has all the data, and you can trace contacts without the knowledge of the users. In EN, the service has a list of infected people, and everyone needs to check that list periodically.
Pretty sure there’s some subtlety with the IDs being a cryptographic sequence or something so there isn’t a gigantic list of IDs everybody is constantly pulling down, but this is the gist of it.
Google have flat out banned any apps that have anything to do with combating COVID-19 if they aren't either funded by their government or are a registered health company. I tried releasing a symptom tracker app before Zoe released theirs and it was rejected for this reason.
Then Google would be building a whitelist. They aren't doing that, they are using a position of power to isolate the market for themselves. That is a monopoly.
No, it doesn’t. They can do the exact thing Apple is doing which is to accept or reject each app on an individual basis instead of blanket refusing every app that is even related to Covid.
What they’re effectively doing is creating a monopoly in the private sector. No other private company is allowed.
Yes, I can. They should do their job of vetting the apps. They have a disclaimer that it will take 7 days then they just reject it for being tangentially related to Covid even if it could have a positive impact.
I was lucky enough to discuss it with google people.
There has been a huge influx of people trying to upload scammy covid apps. As a result the mere mention of covid is enough to get your app flagged automatically.
I don't currently work on a covid app, but I know somebody who does. Their app was flagged but they were able to resolve it by contacting google.
>they were able to resolve it by contacting google.
Do they happen to be government-funded, or represent a health company as stated above? To be fair, I didn't follow up the rejection, I just assumed the rule would be set in stone and they wouldn't budge. The app was free, no ads or any method to profit from covid so I was pretty shocked.
not at all, it couldn't have been smaller : personal app from an unknown dev.
I don't remember the specifics since I thankfully don't get my apps rejected often, but there should be a button to contact the play store support somewhere in the play store UI.
Unfortunately in this kind of situation, the play store handling, while understandable, does not make it easy for legitimate covid apps to be posted.
Somewhat inevitable that the briefing will begin tomorrow against Apple and Google that American big tech firms are deliberately frustrating attempts to fight COVID and wouldn’t it be better if we just taxed them out of existence.
Shame, as a UK citizen I’m entirely supportive of the stance Apple and Google are taking.
I'm not so sure that this type of framing will work so well in the UK. I suspect a lot of the press will jump straight to "it's because Dominic Cummings wants to benefit his mates who do dodgy data mining". Even if that's demonstrably totally untrue, enough people will believe that narrative for it to become mainstream.
The fact that there are plenty of other countries who are content to go with the Google/Apple API will also neutralize a lot of this type of criticism.
I have no idea how to predict the UK public anymore. When it comes to the snoopers charter it went by with barely a whisper from the populace and yet when it comes to the 18+ filter on broadband something like 93% of people had opted out on their home broadband. It's exceptionally rare to find so many people do the opposite of default. I can't work out if the Uk public do or do not care about privacy and their security or what but they surprise me with their actions at times.
"Privacy experts have warned that any cache of location data related to health issues could make businesses and individuals vulnerable to being ostracized if the data is exposed."
Apple and Google already have that data in-house, it's just as vulnerable as any data.
It’s only absurd when you assume that access implies blanket authorization.
Apple Maps is allowed to use my location for the purpose of providing me directions. Sure privacy policies are usually overly broad but for sure “to allow Mark in accounting to track his ex’s” aren’t part of it.
I don’t see a contradiction in two large players refusing to cooperate with governments that want to slurp up people’s location data. They might lose the court case and be forced to allow it but I see no absurdity in the fight.
It's the reasoning they give -- that it might be compromised and used for bad ends. That's already possible, Anyone genuinely thinks that the partitioning of data for different purposes provides any real privacy protections, I've got a LifeLock subscription to sell them.
Lemme guess, libertarian? Or at least "fiscally conservative, socially liberal"? Blind faith in the Free Market Fairy (https://bitworking.org/news/2008/01/The-Free-Market-Fairy) is a big part of how we got here. As Einstein said, "We can't solve problems by using the same kind of thinking we used when we created them".
I can't speak for Google, but I know that Apple goes out of their way not to have any clue where your are. For example, when you do mapping directions, you get issues multiple single-use codes, that are refreshed during the trip, so not only does Apple not know what directions you asked for, they can't even associate the trip with a single entity.
When you say, "Apple has the data in-house" - what are you referring to?
That's assuming the single-use codes are not persisted on the other side. Maybe they aren't.
"Apple has the ability to have the data in-house" / "Apple promises not to have the data in-house" would be more precise. The rest relies on the current behaviour and T&C.
I mean, imagine the repercussions of allowing third party apps to do this. It's tragedy of the commons for health info.
i.e. A crappy mobile app that spams notifications when you're around someone who was in contact with Covid infected. One which doesn't have any oversight and motivation other than mobile ad views.
It means scammy apps will spam contact notifications without proof just to get eyeballs, and that's on top of the ones using covid fears to phish for info. Bad actors will exploit this.
I’m genuinely suprised that an implementation came out so quickly ... one would think the pandemia must have been advanced before the government even considered doing this ... and here we are with Android and iOS approved apps, with bluetooth and permission solutions made by a team of available engineers, implementing goverment documented requirements ... hmmm
There is not a single app using this API in "approved" state, let alone active use. The API is kept very simple, its implementation has just entered a beta stage, and the Bluetooth tech it uses is almost a decade old already. And the entire idea of contact tracing via BT has been discussed for months already by several teams worldwide and already been implemented a few times in various ways, so it's not exactly new either.
But don't worry, just go on with your demonstrative "critical thinking"...
It sounds like this only applies to Google's own exposure notification service and would not apply to standalone contact-tracing apps such as the one being proposed by the UK government.
Its interesting that they call this a "Contact Tracing app" even after changing the naming to ExposureNotification.framework
I think these restrictions are meant to win confidence with a somewhat skeptical public. This will also confine the apps to be single purpose for contact tracing only.
"We have established the person over there has been exposed to Bad Ideas on May 22 2020. Let us find all those people that were exposed to him. We trigger the exposure event via ExposureNotification.framework"
covid19 has become a buzzword factory. Politicians popularize terms like "herd immunity" , "crush the curve", "testing", "ventilators", "PPE" etc to appear to be doing something. "Tracing" is the next in line, but it's a total sham. No country has been able to contain the epidemic with bluetooth. And all the countries that manage the epidemic have first waited until they have very few cases , which can be traced manually, and they did isolation well. As long as there is a high number of active cases, tracing won't work.
So, it's good that apple+google are banning those apps because they would be useless and a damn spying vector.
Schneier's a security expert, not an epidemiology expert. I don't think I'm going to put much weight in his opinion. The UK's epidemiology and behavioural teams have some confidence that this app will have an effect.
It doesn't need to be perfectly effective to be useful. Even a small reduction in R is very helpful.
For what it's worth, I was thinking of developing an app on similar principles a couple of months ago and I talked to at one of the Sage people. They were enthusiastic about it. There have also been papers modelling the effect.
I don't like Cummings' politics but he is a smart guy. I think he'd follow the science.
We took a deeper look in an article linked in another comment on this thread.
One thing that's interesting to think about more deeply is how difficult it is to estimate proximity based on the combinatorial explosion of different hardware, individual device peculiarities, battery levels and environmental factors (walls, glass windows, partitions, ventilation).
The very limited data published appears like interesting preliminary field work which finds significant variability in signal strength across hardware, and rather than concluding proximity estimations are useful, ends in a plea for OEMs to release factory calibration data for their BLE implementations:
Bluetooth is only really an option for Apple, not any iOS developer. Slightly different for Google, where Android lets background apps use Bluetooth. So no country has tried to use bluetooth in their apps. Other than South Korea it sounded like contract tracing had been abandoned.
In South Korea, I suspect that the big difference is the specialist tracing teams they are using combined with the random testing of the public to find new potential sources really quickly. I am sure the app helps but it is probably not the route to normality. It requires thousands of contact tracing teams and high amounts of random testing on very low numbers to work, it is a small piece a combined set of policies that work together.
Would you rather nothing be done? This is potentially a big deal way to automate at scale that has no precedent. It's impossible to say this will/won't work, as it's never been tried.
This is not the opposite of nothing. It's not that phone tracking hasnt been tried, indeed it did not work where it was tried in e.g. in singapore, korea and no other country made it work. What did work, in multiple countries, is "bring the cases down to very low numbers, interview new cases, isolate aggressively". Perhaps they should focus on the latter 2.
> to automate at scale
If you need to track circulating viruses "at scale" you have already failed, as the apps are going to be popping up false positives left and right and you end up with the entire population individually quarantining themselves.
Can anyone show me an app in the wild that uses Bluetooth like or close to what the Covid apps wants to?
Why isn't HN talking about the technical side at all?
We know Bluetooth on phones can't do what the governments says it can.
We've all gone through the stage of, what if we used Bluetooth to track people indoors and do cool stuff! Then we realise you can't. The best we see is advertising maybe doing low quality beacons.
It like we think C19 makes the impossible possible.
The Australian one seems to follow the spec pretty will. Uses rolling random IDs and BT RSSI to check for proximity to infected people. “Infected” is declared by the patients getting the hospital to input a private key when they’re diagnosed, which then uploads their last x days of random IDs to declare them as infected.
Source code is public and has been shared/audited on twitter etc but no formal audits that I’ve seen yet.
> Source code is public and has been shared/audited on twitter etc but no formal audits that I’ve seen yet.
I don't believe that's true. There is certainly decompiled code floating around, but release of the code has been delayed whilst the Signals Directorate investigate the app. [0]
Worth noting that decompiling the app to see if it actually does what it says it does is a crime under the legislation backing it.
> Agreed. The PIA and source code will be released subject to consultation with the Australian Signals Directorate’s Australian Cyber Security Centre.
I've decompiled the app for android and had a look over the source and what you say isn't true.
The rolling ID doesn't work so 3rd parties can track you.
Source code has not been released at all.
It does not work in iOS I think is a fair statement. It won't until Apple do the update.
It does not check the 'proximity to infected people'.
All data processing is done by a human. The app just dumps all your info. All interactions with other phones with the app to a human. Then then work out times and if the person was 'close'
So this is why I've asked the question.
Surely someone on HN has made an android/iOS app and can comment. It would have to be around a phone Bluetoothing to a phone.
Then (in your absurd hypothetical scenario) those agencies are being incredibly stupid, they're wasting their time on on high risk / low reward strategies.
Oh wait, you're probably talking about the United States? It's scary how fucked up that place is just millimetres below the surface.
Where are the bug bounty programs? That's easy. It's a crime to see how any of the app is running. Reporting a security flaw would likely see you receive a $5000 fine, and potential jail time atop of that.
They didn't bother to get the servers running before pushing out a gigantic advertising campaign shaming anyone for not using it.
... Despite it having obvious flaws from day one, that showed it was mostly a cut 'n paste of Singapore's GPL app. (Though you can't access the source. National security trumps freedom of information and promises.)
It isn't actually a law yet - that happens later this month. Instead, we've received a determination by the minister [0], which will act as a kind of back-date for when those laws are passed.
> A person must not decrypt encrypted COVID app data that is stored on a mobile telecommunications device.
That video shows them looking into how the data bundle is assembled, but I don't believe they actually touch it or run it in an emulator, which would very much breach the determination - because unless you're one of the exceptions, you're not legally allowed to run the software outside of tracing.
Exceptions are given for those in employ of the health department, or other government bodies.
Whilst that might vaguely not mean decompiling the app, the minister's own press conference is clearer on the intent [1]:
> It cannot leave the country, it cannot be accessed by anybody other than a state public health official, it cannot be used for any purpose other than the provision of data for the purposes of finding people with whom you have been in close contact, and it is punishable by jail if there is a breach of that.
Decompiling the app steps outside the provisions for looking at the data, and yes, you don't have permission to look at your own data.
IANAL, but I think your interpretation goes beyond what they're after.
The determination is clearly aimed at data usage and prevents people from trying to decrypt the reports from other users. The whole fragment of the interview is about the data produced by the app and how it should be protected as sensitive information. I can't see anything there that would prevent you from reverse engineering "to see how any of the app is running."
It's not even obfuscated or protected from decompilation in any way, so it's trivial to look at with static analysis tools. (i.e. without trying to run it)
Even the headings don't mention the code: "Collection, use or disclosure of COVID app data", "Treatment of COVID app data", "Decrypting COVID app data", "Coercing the use of COVIDSafe".
> IANAL, but I think your interpretation goes beyond what they're after.
Perhaps more than the intent, but this is a government that doesn't deserve the benefit of the doubt.
Circumventing any "access control technical protection measures" is currently a crime under Australian law (Section 116, Copyright Act). They may well consider any decompiling tools to fall under that particular law, as well as use of said tools.
In March, they pressured a university in firing someone researching into their own data breach to see how bad it is. [0] There isn't a law against de-identifying, especially when it is in the public interest, but they went ahead and threatened severe legal action anyway. Whilst simultaneously claiming that said data breach doesn't contain any personally identifiable information.
They had to be taken to the High Court to be shown that an algorithm cannot be used as evidence that a debt exists, and that decision makers actually need to do more than just trust the system. [1]
If it embarrasses them in any way, then they are not above twisting laws to suit them. [2]
So far as I can tell, the legal instrument that sets the requirements for the app is the Biosecurity (Human Biosecurity Emergency) (Human Coronavirus with Pandemic Potential) (Emergency Requirements—Public Health Contact Information) Determination 2020:
567 comments
[ 3.0 ms ] story [ 305 ms ] threadIs the source code of these apps something that could be FOI requested from NHSx seeing as it is publicly funded by the tax payer?
Also they've already started moving the goal posts; https://www.theregister.co.uk/2020/05/04/uk_covid_app_human_...
This* came from NCSC - that image about the NHS version worries me greatly.
* https://www.ncsc.gov.uk/blog-post/security-behind-nhs-contac...
IMHO Apple and Google should have finished the job and "volunteer" to make the app as well as "donate" some of the storages for the data. Not that I trust US companies to respect anyone's privacy, or to the fact that a gag order stapled to a subpoena would give all that data to the "5 eyes"(now 14-15).
I like my health as much as the next living human. But there are so many governments that will jump on this opportunity that it makes me want to avoid this app.
RIPA was passed in the year 2000 under a Labour government when Jack Straw was Home Secretary.
From the independent's website dated 19/11/2016:
The Snooper's Charter passed into law this week – say goodbye to your privacy
https://www.independent.co.uk/voices/snoopers-charter-theres...
They are using the NHS brand to get high adoption.
Plus the UK govt slogan right now is literally "protect the NHS", the existing Conservative party opposition to the NHS is basically gone now, hopefully for good.
One does not follow from the other. Slogans substitute for competency and funding. The opposition is both ideological and directly financially motivated, so I expect the fragmentary privatization to continue.
Watch out for a "now we must pay for coronavirus" user surcharge appearing ...
https://tech.newstatesman.com/coronavirus/palantir-covid19-d...
No, you need viable alternatives for people to switch. Like the Impossible Burger and so on. Until then talking won’t change anything. Not even Snowden level revelations would change it. Sure you can try to use government to fight government. Or... just build the alternative.
We built the Web. We killed AOL MSN and Compuserve.
Build open source, end to end encrypted, self healing and rebalancing networks that use a version of Kademlia DHT that removes IP addresses from each hop. And also run consensus in small groups about stuff.
That’s the future right there. Trouble is, we are only seeing its infancy. I can think of about 2-3 projects that are pulling it off. And they have been working for years.
PS: When this happens, user accounts and quotas will be replaced with crypto, and centralized servers and databases will be only by opt in. They wouldn’t automatically be a thing just because they provide the infrastructure. Infra should become completely commoditized.
PPS: But. You’ll have to worry massivelu about botnets, sybil attacks and massive disinfo campaigns and reputationao attacks by sleeper AI bots. “Fun” times ahead ...
Yup ... I'll take the coronavirus, please.
It’s a giant federated search engine... that’s about it. They don’t even hold any data themselves, it all stays on site with the customer.
If you don’t like the data that organizations record, I’m right there with you, but demonizing the tools that allow them to access data more effectively is an absurd stance to take. Their system also does things like add an immutable audit log, or enforce requirements on how data can be accessed.
Working with law enforcement, before Palantir there was nothing preventing an officer from looking up their ex husband or wife, or celebrities, or anyone at random. With it, there’s an audit trail of every search, and most of them need a case number as justification. Even then, certain searches would automatically get flagged for review.
We can go back to giving people direct and unaudited SQL access if you’d like, but I’d prefer some more accountability, even if the downside is organizations being able to use data _they already have_ more effectively.
But Palantir has been the “tech boogieman” since before I even left in 2014, so it probably always will be. It is admittedly easier to blame a smaller tech company as a scapegoat than try to address the larger governmental or organizational actors that you actually have grievances against.
Just know it won’t do anything. The people who complain online aren’t their target market and never will be. And the paranoia lends way more credence to their dressed up search engine than it really deserves on technical merit alone.
So in a way the people complaining about it are making the issue they’re so upset about worse.
Or you can do both. You might as well suppose that people are critical of the mercenary company Blackwater because they want to avoid criticizing America foreign policy and military exploits. I don't think that's accurate at all, it's been my experience that people who are critical of one are very often critical of the other as well.
Similarly, it's been my experience that people critical of Palintar are also critical of the organizations and governments who use the services of Palintar.
And yes, of course, complaining about Palantir online isn't going to change the government policy. But if working for Palantir means that no other self-respecting software engineer will want to shake your hand, then fewer people will want to work there, and their surveillance tech will be lower quality and have more holes in it.
If so, please don’t tell me you’re a DBA.
A crucial step in shooting someone (or a drone strike) is figuring out who to target and where they are.
The fact that this is an important use case for Palantir makes it not hyperbole to talk about "weaponising information".
Everything has a domain it operates in, so I believe that statement has nuance. In particular when you look at how general or specific the domain is.
From the perspective of something uploaded to S3? It sure is like any other file.
From the perspective of a website dedicated to the dissemination of child porn? It clearly is more than "just bits" in that context.
How about from the perspective of a search engine, where that site may be indexed? Welcome to the grey area that all of these debates are rooted in. Technically it's just indexed strings or ngrams, so it is like any other type of file. But there's an argument that a search engine should "know more" than just the raw data, and should be able to understand that context somehow.
Palantir is in this grey area. The software isn't built for spying. It's built for managing and understanding vast amounts of data. Can this be used for spying? Yes. It can also be used for double blind clinical trials. Or maintaining insurance systems. Or coordinating disaster relief efforts.
It operates technologically in a very general domain, but their flexible user-defined ontology makes it very powerful at operating in more specific domains. So it's a lot less cut and dry than the dissenters make it seem, IMHO.
I don't think everyone associates government agencies with pristine records of holding themselves accountable.
If you’ve never worked with a government agency before, let me tell you, Hanlon’s Razor is in full effect. Never attribute to malice that which is adequately explained by stupidity.
The higher ups know that free reign for their officers to look up anything is a recipe for legal nightmares. The individual officers are often just too dumb to realize that they shouldn’t search for something, and do so out of curiosity. Having blocks in the way like requiring a case number be entered to run a search, and knowing that that search and results are forever associated to that case, cut down a lot on abuse.
Are all forms of abuse the same, or so easily mitigated? No, but I’d like for some system to audit that and at least try to enforce it.
The irony to me is that when I was at Palantir they talked extremely openly about all their technology. A ton of it was open sourced, and in depth tech talks were posted to YouTube, but almost no one watched them.
I think people just like writing them off as shady or sketchy because it’s easier than needing to reason about the clear ethical nuance that exists in the government contractor space.
You know, my dad always told me to just listen to people... eventually they'll tell you who they truly are.
The hits were all more or less expected, a bunch around their home and the precinct, local supermarkets, etc. But then a weird grouping way out of town. Detective insisted it was a mistake and wanted more data, thinking someone stole his license plate (insert eye roll here).
Anyways, zooming in and looking at the highest density of hits, it ended up being a strip club that he frequented. He got bright red and his buddies didn't stop giving him shit for it all day.
And this was a detective, not a beat officer.
I’m not saying it’s perfect, but scapegoating the company is avoiding the real issue, which is the government policies or the enactment thereof.
HN won't let me reply to that, so I'll do it here. I didn't feel the need to offer a "defense" of Palantir there since I already outlined it above and don't feel like there's value in repeating myself.
But you asked, here you go - The government is going to use the data regardless, so long as they have it. The only effective recourse is to make them collecting or storing the data illegal, but good luck with that, seeing as even when it is illegal they do it anyways (see: Snowden).
Do you think is Palantir weren't there they'd just say "Ah well, I guess that's that" and be done with it? Nope; they'd go to Lockheed, Raytheon, IBM, or another long-time contractor to build a replacement almost immediately. The technology is genuinely nothing special. It federates searches across disparate database, and stitches results together.
But for me, having worked at Palantir, I know what mechanisms are provided to attempt to mitigate abuse. So as long as that data does exist and is going to be used, I'd prefer to have it be as well controlled and audited as possible. And sure, you can argue that the government may forego auditing, or be entirely corrupt, but that doesn't seem to me to be a good reason to not use tools that at least provide that capability.
If you want to see change that is more than just superficial, that’s where you need to make it.
Edit: Changed "we don't" to "I don't". I don't want to claim to speak for anyone else but myself.
I'm talking about neutering the surveillance state. That's not something that demonizing individual companies is ever going to do. You need to push for that change at the government level.
Unless you're pro-surveillance, but your previous comments made it seem like the opposite.
If someone comes into a gun shop and tells you they want to shoot up a classroom, selling them a pistol instead of a machine gun doesn't make you immune from judgement.
Want to know how a lot of that data is generated? Microsoft excel. Arguably without excel there would be less data to act abusively with. If only Microsoft just refused to sell excel to the government.
People are happy to accept that excel is general enough that it isn’t made for that one purpose, but refuse to apply the same reasoning to Palantir; that may seem like whataboutism, but they genuinely are comparable tools if you take away the marketing and fear mongering.
As for your analogy, it’s more like you own shop that makes metalworking tools, and someone buys some. You don’t know what they are going to do with it, and (here’s where it gets opinionated) you shouldn’t need to care. You sold a tool to someone. Can the tool be used to make a gun? Sure. But it can also be used for anything else involving metal work. At some point we need to accept that the responsibility for how a tool of used needs to be placed on the person using it.
We do not hear, "Government X purchases a couple thousand Excel licenses to keep a database on people it'd like to kill", and I imagine that Microsoft does not sell Excel to Government X in response to an RFP for tooling to keep track of people it'd like to kill.
(Of course, it'd probably stick its metaphorical fingers in its ears if it did hear about Excel being used for that purpose.)
So, basically "screw rule of law, people are gonna do it anyway, so let's make money off it!"
Do you ever use uber? They work to keep their drivers legally declared as contractors, despite treating them employees in all other regards. That’s pretty immoral, so I hope you don’t encourage it by supplying them with business.
Machine Learning/AI adds an entirely new perspective on big data. It can see patterns in data that would be meaningless without it, correlations that can't be found in other ways. Some data is OK for a government to have if they can just look it up in exceptional cases, but not if they act on every single bit of data collected.
In essence, using ML is creating new data (or at least: Information from the raw data). This information can be very revealing and not at all in line with the scope the data was initially collected under, which is one of the principles of GDPR.
You can't really view data as static once it's collected. Combining with other existing data and new methods of data analysis make it a totally different picture in terms of privacy.
I’d love to be in the utopia world where there is no war and the entire planet gets along. But even the US has political bipolar disorder, so we’re a looooong way from global peace.
I do think we spend an absurd and inappropriate amount on defense, but as long as we are forced to develop weapons due to the global geopolitical climate, I’d consider it preferable to work on making them more targeted.
My problem is I don’t like or agree with the people picking the targets, but that’s a whole other argument.
We could absolutely stop all weapons production in the US, I’d just ask for some time to learn Mandarin first.
The work defence contractors do and will continue to do is a measurable positive in the world. This work generally falls into three categories:
First is projects that make the lives of armed forces personnel easier. There is no moral or ethical hazard with these projects. They simply are making it so that pre-existing work, often non-combat work, is less tedious.
Second is projects that actively protect the lives of armed forces personnel and assets. Why would somebody have any reason to feel anything but pride in their work when they know that the only things it could ever do is save lives.
Third is the arguably morally shaky stuff. Weapons technology would fall under this category. The thing people seem to miss about this category however is what new projects' goals are. Very rarely is it ever to create a more lethal weapon. Almost all development of this kind focuses on either making existing technology cheaper, more precise, or have increased range.
In the first case, you could argue that it makes the weapons more likely to be used which is fair however if a weapon system is in active use already, it is unlikely that a cheaper version would do anything other than bring down operating costs. In the second case, these projects are actively making these weapons safer. Look at the Hellfire R9X as a prime example. This weapon sole purpose is to minimise casualties and minimise damage to the surrounding area. Finally the third and last case I mentioned. Increasing range is important for force projection reasons but the most immediate benefit is that it allows armed forces personnel to be further from danger. These people would be in this fight regardless but now they are able to operate from a safer distance instead.
Weapons tech gets a bad rap for obvious reasons but I see that as mostly having been a relic of a past era. Nowadays there isn't really any reason to make more lethal weapons. We already have those and you can see a steady trend in new military technology towards lower risks, costs, damage to the surroundings, and minimising civilian casualties.
Additionally, in my experience, you generally have a choice whether you work on a project or not. Leadership fully understands what they do as an organisation and get that some people may not be comfortable with doing certain projects. If you can't conscientiously work on a certain type of project, simply say so and leadership will take that into account when planning personnel assignments for projects.
TL;DR Armed conflict and existing weapons will never go away. What we as a society(namely the modern purpose of defence contractors) can do instead is make armed conflict less dangerous to those caught in the crossfire and those who risk life and limb to defend our countries.
Not judging your decision (although I personally prefer not working for defense/defense contractors), just pointing out that "it only makes soldiers safer!" isn't an ironclad rebuttal.
Analogy: Would you consider it unethical to provide tools and services that make the lives of human traffickers easier? To actively protect assets of drug lords distributing contaminated opiates and meth?
Slave- and drug trade are about as likely to go away as armed conflict.
This is in the same way that corrupt charity organisations can defraud the causes they claim to support and corporations can knowingly risk the lives of their workers due to either negligence, mismanagement, or for the sole purpose of saving costs or making more money. This doesn't make charities or corporations unethical solely because they have the potential to be.
Additionally, armed forces can and do maintain their operational fitness by using their immense logistics networks and pools of human labour to do very real quantifiable good in the world. Outside of their use as a deterrent and show of power, the US armed forces spends a large chunk of their time and effort responding to natural disasters, taking part in humanitarian efforts, and building infrastructure.
---
On to the analogy, I would like to reframe both examples a bit.
For the purpose of this discussion, illegal transportation of individuals falls into two categories: Human trafficking and human smuggling. Human trafficking is done without consent of the individuals being transported. Conversely, human smuggling is done with consent. Despite being illegal, human smuggling provides a very important service to individuals. It may be against the law but it can be argued that it can provide a net benefit.
Mind you that it comes with severe downsides and by no means am I trying to minimise those. I'm just trying to keep this from turning into a 20 page essay on the topic. One of those downsides namely is that it increases the volume of people being transported which makes it cheaper and easier for human trafficking rings to operate. Another downside is that a large number of people die during transportation, both those consenting and non-consenting.
Now if you move onto the less clear cut outcomes, human smuggling allows individuals to enter a country without going through the vetting process. This opens a vector for criminals and terrorists to enter a country however at the same time, it provides a means for people who either aren't willing or aren't able to wait untold years to get a visa to enter the country only for it to then not be renewed a few months or years down the road. These may be individuals or families that want to be productive members of society in a country with a higher standard of living or are trying to move to a location where they can be treated for medical issues that they can't be treated for at home. Another common case in human smuggling is individuals using human smuggling to cross countries that have closed their borders to them so that they can get to a country that will accept them as refugees. The number of reasons somebody would illegally enter a country go on quite a bit and the list is full of both morally sound and morally reprehensible reasons.
With that in mind, while I personally wouldn't participate in any type of business connected to the illegal movement of people across national borders, I see no moral or ethical issues with someone who does so for the sole purpose of making the process safer. Despite the fact that it is illegal, the way I see it is that you are reducing the risk that people lose life and limb.
---
Moving on to the other example. My opinion is that most drugs should be legalised and regulated. By moving to a legitimate market and starving these organisations' source of income, we should see drug lords be replaced by corporations. You could argue that the cartels would jus...
> These people would be in this fight regardless but now they are able to operate from a safer distance instead.
I sincerely strongly doubt US would be fighting 5 wars with boots-on-the-ground in Afganistan, Pakistan, Somalia, Yemen and Iraq, if it weren't for remotely-operated drones.
However, I also believe you are responsible for the consequences of the technology you create. If you decide to work on smart bombs and because of a bad government those end up killing innocent civilians, some of that blood is on your hands. If you're willing to accept the risk of that, I respect your choice.
Personally, I would rather avoid that by not working on that category of technology at all. It's not the kind of mark I want to leave on the world. I do appreciate that I only have the luxury of this choice because I live in a country where others do choose to work on defense so that I can be protected.
I agree with this to an extent, but at the same time, technologies you develop can go well beyond what you intended them to be. Do you think Ritchie knew C would be used the way it is today? It is used in UAVs, smart missiles, and more. Building faster algorithms also means faster weapons. Everyone working at SpaceX is directly or indirectly working on missile technology. You can do this with practically any technology, even as simple as building a better screw. Building materials that are stronger, lighter, and cheaper reduce the prices of homes, but it also has applications in war. Studying diseases and vaccines directly impacts biological warfare research.
It is easy to say you're responsible when you're working on things like smart bombs. But it isn't easy if you're researching fertilizer. The thing that arguably has saved the most lives yet how many lives have bombs and bullets taken?
> Personally, I would rather avoid that by not working on that category of technology at all.
So what I'm saying is you're working on that tech, just how removed are you? I also don't think there's a cognitive dissonance. You can be pro putting nitrogen in soil and anti explosive nitrates. But saying the two aren't related is naive.
Engaging in any kind of commerce (i.e. beyond growing food for yourself and your family) means paying taxes, which means funding governments and armies.
I don’t think both of those can be true simultaneously, unless you somehow consider “having blood on your hands” to be ethical.
If you're working on weapon systems that (you know or expect) will be just be directly used for murdering[1] innocent civilians, then you're endorsing murdering innocent civilians.
If you don't know where on that scale things are, then you're reasoning under uncertainty, and your ethics are going to have to deal with that.
0: technically most of those would actually be manslaughter, going by the usual definitions, but a: not all, b: that's not a verb.
1: nope, just murder
I don't generally believe in black and white, so I'm making no claims that anything is 100% ethical, 100% unethical, 100% blood on your hands, etc. I think it's possible for good people to work on weapons and still sleep soundly at night. I also think it's possible for good people to work on weapons and end up regretting the consequences of that choice.
People pretend they can't possibly know how things will get used, but this has never been true, and never been less true than today.
http://tech.mit.edu/V105/N16/weisen.16n.html
It's rather amazing the mental gymnastics people perform to get out of cognitive dissonance. I'm much more sympathetic with people who know exactly what they're working on and the full effects of that work, why some will find it distasteful, but nonetheless can at least make a case for why those people are wrong and the work important, rather than agreeing with them but hastily adding some transparent excuse to avoid being lumped in the distasteful category.
If you don’t like the bullshit the government does, that seems like an issue you should take up with the government, no? You’re using a tech company as a scapegoat.
And that's the problem – they're giving governments tools that let them more easily and effectively infringe on the privacy rights of their citizens. Privacy isn't invaded when the data is collected, although that's a necessary step; it's invaded when a cop or bureaucrat queries that data to see what a citizen has done. Giving them that tool is morally complicit.
~ "That's not my department," said Wernher von Braun. ~
Quite frankly, I don't think any reasonable person would consider selling software to the government to be morally corrupt unless they were reasonably confident the government would use it for evil. Right now, nobody has made a compelling-enough argument to make me believe the government will use this software for evil.
As for compelling arguments... is the history of most of the world's governments not enough for you? In US, you can look at the census for a case in point: this data was used to chase draft dodgers during WW1, and to compile list of Japanese for internment during WW2. Curiously, by WW2, the federal census law had specific provisions preventing the Bureau from disclosing that information to any other government agencies, precisely to prevent this kind of use - Congress simply repealed those provisions. Nevertheless, the Bureau subsequently denied sharing that data and buried any leads, so we didn't have definitive proof until this century.
Or we could talk about COINTELPRO, PRISM etc.
There are popular licenses with such clauses.
This is software that will ostensibly cut down on errors and speed up processing times for things like asylum claims, but for some reason Github was called out as being complicit.
I think the argument against Palantir is better formed than an argument against Github based on specificity of the tool, but I think both are driven more by viral outrage than careful consideration of any moral calculus or actual hands-on knowledge of how the tools are used.
Palantir is _explicitly_ contracted by a government to create a tool for combining surveillance information to make tracking individuals more effective.
Github is providing tools (Github Enterprise IIRC) which is presumable being used to create tools which are used to track and deport individuals.
Other databases get used to track down, torture, and murder critical journalists - by dismembering them with a bone saw while they're still alive.
Remind me again which kind of database Palantir staff work on?
The software was also used to run double blind clinical drug trials - the acl granularity was great at that; you could have different roles for drug manufacturers, doctors, and patients, none of which had absolute information. Then trial supervisors could open up the data for analysis at the end of the trial.
At that same time, the ability to use GPS phones to upload data over unreliable networks was used when they teamed up with the Clinton foundation and Team Rubicon to improve disaster relief coordination for hurricanes.
Myself or coworkers (“Palantir staff”) worked on all of these.
Also, I'd like to point out that this isn't an "either or" situation. Obviously the government is involved, as they no doubt put out the contracts, but Palantir is the one cashing the check. They've created their entire business model around big secretive government contracts.
Holy shit, and I didn't think that this whole contact tracing bullshit could look even worse.
I mean besides “zero to one” and “competition is for losers, build a monopoly”?
I wonder how much effect he had on Zuck early on. He was the first big check in. Seems it was a great match there.
Principles dont matter if you dont have the money to put them into practice.
If palantir is what it takes for him to fund/back a visionary company or leader that will create a truly freer society, then that investment was well worth it.
My question to you is why you care what Thiel thinks and why you're dissapointed. As far as I'm concerned, these guys are the lowest of the low and they only hold their position through gross exploitation.
Nobody should be amassing this impossible wealth and seeking to gain more and these CEOs are basically economic despots to my mind.
"In a covenant concluded among proprietor and community tenants for the purpose of protecting their private property, no such thing as a right to free (unlimited) speech exists, not even to unlimited speech on one's own tenant-property. One may say innumerable things and promote almost any idea under the sun, but naturally no one is permitted to advocate ideas contrary to the very purpose of the covenant of preserving and protecting private property, such as democracy and communism. There can be no tolerance toward democrats and communists in a libertarian social order. They will have to be physically separated and expelled from society. Likewise, in a covenant founded for the purpose of protecting family and kin, there can be no tolerance toward those habitually promoting lifestyles incompatible with this goal. They – the advocates of alternative, non-family and kin-centered lifestyles such as, for instance, individual hedonism, parasitism, nature-environment worship, homosexuality, or communism – will have to be physically removed from society, too, if one is to maintain a libertarian order."
This quote is from Hoppe's "Democracy: The God That Failed", which Thiel has referenced before. The overall thesis there is that economic freedom is the cornerstone of libertarianism, and it's fundamentally incompatible with democracy, so libertarians have to explore other options. It specifically promotes monarchy, on the basis that a monarch is more like a business owner with a vested long-term interest.
You might argue that this isn't really libertarianism - indeed, many libertarians don't consider Hoppe to be one. This is basically proto-neo-reaction, at the point where it still hasn't explicitly rejected libertarianism altogether. But regardless of where you draw the line, this all is the evolution of Murray Rothbard's strain of libertarianism, after it moved away from social liberalism towards paleoconservatism - it's not just some random people coming and declaring their politics to be "libertarian" out of the blue. Nor is there any shortage of ex-libertarians in NRx circles, so it's clearly a common path, not unique to Thiel.
Needless to say, the libertarian paradise described above could use something like Palantir to implement the "covenant".
But the label is irrelevant, Thiel and Palantir's actions are anti-democratic and their involvement in any government activities immediately raise privacy and safety concerns.
Do you have any references showing Thiel believes these words you quoted?
In general, Thiel's political opinions are hardly a secret; you can read his own words at https://www.cato-unbound.org/2009/04/13/peter-thiel/educatio...:
"I no longer believe that freedom and democracy are compatible"
This is literally the primary thesis of Hoppe's book.
And note that Hoppe's isn't saying that promoting homosexuality should get one expelled from society, but rather from communities with "covenants" that restrict it. This arrangement is all about one-dollar-one-vote, so rich people needn't worry - they can live however they want, even establishing their own "covenant" if need be.
That explanation is much more reasonable. If people can choose which covenant to belong to, that's better then some larger collective (like a federal government) imposing it's morals on smaller groups.
None. He is a slimy character with no principles or morals besides his own selfish drive. He wants to make money, period. Other considerations don't factor into it. Make no mistake, he (and people like him, the capitalist sociopaths that hold power in our society) would kill your entire family for a handful of pennies if they would get away with it.
If there were any justice in the world these kinds of people would be shunned and censored by society. Yet it's precisely the opposite. These are the behaviours which are encouraged and rewarded.
The issue is that the UK govt is composed of a combination of incompetent and corrupt people (one of the companies involved in this is owned by a minister). And they are trying to go with the low-cost option (whilst the govt is subsidising 1/4 of salaries paid in the whole country) rather than the old-school, proven way.
It is kind of shocking but it is the result of a political system (and culture) that rewards incompetence and sloth.
That's not untrue, but that's clearly not contact tracing, much less was-in-the-same-room tracing.
The reason I ask is, I also thought that it's "just supreme evil" until I listened to a podcast with Thiel (Dave Rubin podcast) recently... In it, he said that he started Palantir as an experiment, what can be done against organised crime/terrorism while respecting civil liberties and privacy. His argument was, that we do need to stop terrorism, because people (not HN crowd, but the public at large) are obviously not willing to compromise safety for privacy, so each terrorist attack results in privacy- & civil liberties-curtailing laws - so the question is, can we use data/ML/etc. to safeguard our society while preserving as much privacy & liberty as we can.
But I've almost no real (and insider) info about Palantir to judge whether that's actually what the organisation is, 10 years or so later.
Having said that I know nothing of Palantir, today was the first time I heard of it.
But I would never ever believe anything a CEO says in the media that can't be validated. Their very job is setting the company in a positive light above all else. Especially when it comes to vague terms about company ideals that can't be proven. And 10 years is an eternity in IT. Remember Google had their motto "Don't be evil". Look how that turned out.
There's no need for such a request. They themselves say "We intend to open source our codebase once the design is finalised" here: https://www.ncsc.gov.uk/report/nhs-covid-19-app-privacy-secu...
https://www.bbc.com/news/explainers-52442754
That's what they recently told us they were not going to do after all. Because security.
:sigh:
I don't even think it will even be effective given the anticipated number of cases.
A contact tracing system and a virus with n days of non-symptomatic incubation, and international borders. Let's think this through a bit. Of what benefit is to the traveler admitting/host country (H) to note on the "medical passport" issued by origin country (O)? Obviously, for this thing to even work on paper, countries need to exchange massive amount of information.
The fact stands [challenge it!] that the socio-political regime being rolled out in guise of defense against Covid-19 and future viral friends aligns perfectly with a specific ideological position. It really was 'Nature's gift'.
What used to be accomplished via violent revolution is being accomplished with 'scientific' efficiency and very little blood. Oh, that figurative blood in the streets of middle and lower classes? "The Virus does not discriminate!".
I'm not sure about the hand me down binary model conclusion. We're in agreement regarding the inevitability of this turn of events, but then again it was apparently clear as day thousands of years ago to "self isolating" monks.
But the clarity of the choice is stark, and like a blast of cold air invigorates one's morals and character.
Good luck!
In a way I think this will actually hamper their efforts.
After all: With the corona tracing apps, people are becoming much more aware that their data is used to trace them. A lot of people I know are talking about leaving their phones at home when they go out.
Considering the carriers have had this data forever and are probably sharing it, this awareness is a good thing for privacy.
I've worked at a telco where we were analysing customer's behaviours and almost everyone follows a pattern at some point e.g. going from home, via public transport to work. The resolution isn't high but over time you build up enough samples to accurately determine exactly where they live, work etc.
You need to have enough people for that, you need to have the testing capacity and you need to have people following that order.
It will work well if you have very low community spread and can start tracing on cluster level. If you are not there it can still be a good tool but you would probably still need a lot of other restrictions as well.
If we do end up implementing a centralized contact tracing system, the results will be very interesting. I would love to know the name and location of every lobbyist that my elected representatives have met with, and I'm sure that in the name of public health they won't make any efforts to thwart this tracking.
In that case the governments do not need this silly contact tracing app. They can just use the data they already have.
We were required to sell the data to the government on demand due to CALEA and similar legislation.
Those who complain endlessly about AANG largely seem to have no idea of how our communications networks actually work. Your location, activities, banking, etc. are not private, will never be private, have never been private, and it has nothing to do with AANG. The 3 AANG that I have worked are woolly lambs. That F guy is a bit suspicious, but I don't have first hand knowledge.
For a start, installation of the app is voluntary so if you're planning something you really don't want the government to discover, you wouldn't install it.
Also, intelligence services already get call metadata so they know which mobile towers you are using and probably lots else.
As for effectiveness, I think it will be useful. There have been several papers crunching the numbers and it looks like it will have an effect. It doesn't need to be perfect - any non-trivial reduction in R is helpful.
https://www.hsj.co.uk/technology-and-innovation/exclusive-wo...
If this was written by NSHX I would install it, but given the dodgy provenance and lack of open source there's no way I'm going to install it.
"We intend to open source our codebase once the first release is finalised. The documentation accompanying that release will supersede this paper."
and from https://www.ncsc.gov.uk/blog-post/security-behind-nhs-contac...,
"The Secretary of State has also committed to making the source code open source. That may not happen immediately on release because the fabulous NHSX development team are all working hard on getting the product ready. But it will happen."
Sounds like there is an intention to deliver on the source. Bigger question is whether the server side is open sourced (as clearly that has limited value for verifying what is happening remotely).
That's just annoying nonsense though isn't it?
There are reasonable (well 'ok they happen, whatever') causes for delay in open-sourcing something that didn't start out that way, but... Do any of them take more than seconds of developer time? They mean it 'may not happen immediately, because it's stuck in legal', surely?
Also while the app on the phone has been released, the actual data wont be sent on to contract tracers until more testing and privacy policies are finalised.[0] [0]https://www.abc.net.au/news/2020-05-02/coronavirus-app-curre...
The source code should be released before the public release of the binaries, not after, and not doing so means they're trying to hide what the app does from timely public discourse.
Releasing a tarball (even if it lacks the tools to build it) isn't that hard, is it?
https://github.com/nhsx/COVID-19-app-Android-BETA
https://github.com/nhsx/COVID-19-app-iOS-BETA
The most important thing is to get absolutely as many people possible using the technology.
Besides, for their own selfish reasons, neither Google or Apple want to be so blatantly attached to something that could be so easily abused.
Google and Apple need to advocate privacy for reasons. But they also want to help the government (PRISM example). So they carefully craft a bug that only government knows about and that allows the extraction of location data. If ever discovered it was not intentional. The NSA backdoor at RSA is an example [1].
[1] https://www.reuters.com/article/us-usa-security-nsa-rsa/excl...
Previously Apple were made to bend their rules when India threatened to ban Apple devices if they don't allow TRAI Do Not Disturb app in 2018.[2]
[1] https://play.google.com/store/apps/details?id=nic.goi.aarogy...
[2] https://9to5mac.com/2018/11/30/apple-approves-india-dnd-app/
[3] https://paste.gg/p/anonymous/b7c95d3967514e78a652840b5b666d5...
It's scheduled for "mid-May"
Also I personally think the permissions(location and bluetooth) are fine for an app like this to really function. I have read someone mentioned on HN that these platforms prove their worth when >60% people are using them(I maybe remembering wrong though).
I do that will all the apps that require location access: local food delivery, cab services, vehicle rental and what not.
Recently this has also been made mandatory for employees, public and private. So organizations have to ensure all employees have this app on their smartphones. We will see how much this is enforced.
App knows their location, and app knows you’re near them. Location access by proxy.
Being on a stock, unrooted phone, the thing I miss most is xprivacy's prompts whenever an app wants to use a permission. I just make sure to check my permissions list every month or so to make sure Google hasn't silently allowed an app update to enable new permissions.
And if you disable iCloud, the data remains exclusively on your device.
And if you're in China, Apple stores both on Chinese government owned servers: https://www.amnesty.org/en/latest/news/2018/02/5-things-you-...
To view Significant Locations, I have to provide my Touch ID again. Also, the fine print says "Significant Locations are end-to-end encrypted and cannot be read by Apple." Text repeated in "Location Services & Privacy"[0].
[0] https://support.apple.com/en-us/HT207056
> It is used to provide you with personalized services, such as predictive traffic routing, and to build better Memories in Photos.
That's not Apple keeping the data, that's your phone keeping the data. And that data isn't accessible to Apple - only you and your device.
From page 6 of https://www.apple.com/privacy/docs/Location_Services_White_P...
"This data is not shared with third parties, is fully encrypted, and can’t be read by Apple."
Yes, the Equifax incident really showed us how data leaks ruin companies.
Sure, neither is probably perfect, but only one of them is an egregious violator whose business model depends on surveillance capitalism
It should be a complete no-brainer that everyone who is ok using google or apple maps, i.e. probably 95+% of the country, be ok downloading their local health department's contact tracing app if it uses these APIs. But we'll see how it plays out, in the end it will surely come down to how this is politicized and not actually be based on the technical merits of this protocol at all.
Free GPS and navigation services have a price. I doubt many people are going back to Garmin anytime soon. Just the way it is now.
there are other options though. ive been using openstreetmap for about 4 years and have only found the need to use Google once in all that time. although with openstreetmap being crowd sources it really depends on how much work has been done in a particular area but in my country its pretty decent
1. It’s actually crap and they don’t want their advertisers to know that their ads are sold based on crap quality data/don’t want the backlash of doing a poor job helping
2. Most people at google care a lot about privacy (either in the secret kept between you and google sense or the more common sense definition people seem to use on hn) and they don’t really think about this data as something the firm has/should release
3. They are afraid that if governments realised they had this data then google would be regulated or every minor security agency/random government department would be demanding access to it by law.
4. They strongly feel that surveillance by (well intentioned?) private companies is ok but by governments it os not.
When I worked at Google, I saw a lot of earnest efforts to keep data private, in a way that really was sufficient (k-anonymized with large k, for example), but in ways that have no outward proof that it was happening. Send all possible data to the server, then make sure it's properly clustered and scrubbed before it gets stored or analyzed. And it's not easy to explain to someone who can see the whole system that from an end-user's view this is identical to just scooping up everything.
5. They very much want this project to succeed so that it can save lives. They know privacy concerns are (rightly) an existential threat to the success of the project, so they are trying to address those by drawing this firm line.
It's not 1 or 3 because anyone can look at the location data Google gathers on them and judge for themselves if it's crap. Just search for "view google location history" and it will take you right there.
I think 2 is pretty close to the mark. Despite HN's understandably cynical take, everyone I talk to at Google cares a lot about user privacy.
You can only see the tiny fraction of data that relates to you, and only the data that has been identified to you. This says very little about the quality of the rest of the data google have. Doing the double slit experiment once with a single election will not tell you much about quantum mechanics.
is that really all there is? they are showing you everything? theres no shadow profile?
Also, Google and Apple where part of PRISM. So there is a history of cooperation.
Nothing prevents anyone from using their phone's location history to remember what to tell the contact tracing people.
Contact tracing: Alice is able to say, “I was in contact with Bill and Carol.” Then authorities can talk to Bill and Carol, and have them trigger their phones to see who they’ve been near. But because that’s slow, most plans would upload the lists of who’s been near who to a central server. Then the authorities can do a simple query to see who’s been near who.
Exposure notification: Alice enters a code that she got with her positive test result in to the app. The app has been continually broadcasting rotating, random identifiers which it then uploads to the central service. The code she entered verifies to the central service that she has a legitimate positive result. Bob and Carol’s phones periodically check with the central server for the list of positive IDs. Their phones stored one of the IDs from Alice’s phone when they were near each other earlier. Once they get the latest list of infected IDs, their phones will alert them that they have been exposed and should be tested.
In CT, the central service has all the data, and you can trace contacts without the knowledge of the users. In EN, the service has a list of infected people, and everyone needs to check that list periodically.
Pretty sure there’s some subtlety with the IDs being a cryptographic sequence or something so there isn’t a gigantic list of IDs everybody is constantly pulling down, but this is the gist of it.
ETA: The FAQ from Apple+Google is a pretty quick rundown of where exactly each part of the data is stored and when it leaves your device. https://blog.google/documents/73/Exposure_Notification_-_FAQ...
It's a lot easier for Google to operate on a whitelist model than a blacklist model.
What they’re effectively doing is creating a monopoly in the private sector. No other private company is allowed.
There has been a huge influx of people trying to upload scammy covid apps. As a result the mere mention of covid is enough to get your app flagged automatically.
I don't currently work on a covid app, but I know somebody who does. Their app was flagged but they were able to resolve it by contacting google.
Do they happen to be government-funded, or represent a health company as stated above? To be fair, I didn't follow up the rejection, I just assumed the rule would be set in stone and they wouldn't budge. The app was free, no ads or any method to profit from covid so I was pretty shocked.
I don't remember the specifics since I thankfully don't get my apps rejected often, but there should be a button to contact the play store support somewhere in the play store UI.
Unfortunately in this kind of situation, the play store handling, while understandable, does not make it easy for legitimate covid apps to be posted.
Shame, as a UK citizen I’m entirely supportive of the stance Apple and Google are taking.
The fact that there are plenty of other countries who are content to go with the Google/Apple API will also neutralize a lot of this type of criticism.
"Privacy experts have warned that any cache of location data related to health issues could make businesses and individuals vulnerable to being ostracized if the data is exposed."
Apple and Google already have that data in-house, it's just as vulnerable as any data.
How is this not an absurd stance?
Apple Maps is allowed to use my location for the purpose of providing me directions. Sure privacy policies are usually overly broad but for sure “to allow Mark in accounting to track his ex’s” aren’t part of it.
I don’t see a contradiction in two large players refusing to cooperate with governments that want to slurp up people’s location data. They might lose the court case and be forced to allow it but I see no absurdity in the fight.
When you say, "Apple has the data in-house" - what are you referring to?
That's assuming the single-use codes are not persisted on the other side. Maybe they aren't.
"Apple has the ability to have the data in-house" / "Apple promises not to have the data in-house" would be more precise. The rest relies on the current behaviour and T&C.
i.e. A crappy mobile app that spams notifications when you're around someone who was in contact with Covid infected. One which doesn't have any oversight and motivation other than mobile ad views.
But don't worry, just go on with your demonstrative "critical thinking"...
https://blog.google/documents/72/Exposure_Notifications_Serv...
Section 3.c.i
It sounds like this only applies to Google's own exposure notification service and would not apply to standalone contact-tracing apps such as the one being proposed by the UK government.
I think these restrictions are meant to win confidence with a somewhat skeptical public. This will also confine the apps to be single purpose for contact tracing only.
So, it's good that apple+google are banning those apps because they would be useless and a damn spying vector.
It doesn't need to be perfectly effective to be useful. Even a small reduction in R is very helpful.
For all we know this is Cummings not understanding any of the science.
I don't like Cummings' politics but he is a smart guy. I think he'd follow the science.
One thing that's interesting to think about more deeply is how difficult it is to estimate proximity based on the combinatorial explosion of different hardware, individual device peculiarities, battery levels and environmental factors (walls, glass windows, partitions, ventilation).
The very limited data published appears like interesting preliminary field work which finds significant variability in signal strength across hardware, and rather than concluding proximity estimations are useful, ends in a plea for OEMs to release factory calibration data for their BLE implementations:
https://github.com/opentrace-community/opentrace-calibration...
Perfect is the enemy of progress.
> to automate at scale
If you need to track circulating viruses "at scale" you have already failed, as the apps are going to be popping up false positives left and right and you end up with the entire population individually quarantining themselves.
Why isn't HN talking about the technical side at all?
We know Bluetooth on phones can't do what the governments says it can.
We've all gone through the stage of, what if we used Bluetooth to track people indoors and do cool stuff! Then we realise you can't. The best we see is advertising maybe doing low quality beacons.
It like we think C19 makes the impossible possible.
Source code is public and has been shared/audited on twitter etc but no formal audits that I’ve seen yet.
I don't believe that's true. There is certainly decompiled code floating around, but release of the code has been delayed whilst the Signals Directorate investigate the app. [0]
Worth noting that decompiling the app to see if it actually does what it says it does is a crime under the legislation backing it.
> Agreed. The PIA and source code will be released subject to consultation with the Australian Signals Directorate’s Australian Cyber Security Centre.
[0] https://www.health.gov.au/sites/default/files/documents/2020...
The rolling ID doesn't work so 3rd parties can track you.
Source code has not been released at all.
It does not work in iOS I think is a fair statement. It won't until Apple do the update.
It does not check the 'proximity to infected people'.
All data processing is done by a human. The app just dumps all your info. All interactions with other phones with the app to a human. Then then work out times and if the person was 'close'
So this is why I've asked the question.
Surely someone on HN has made an android/iOS app and can comment. It would have to be around a phone Bluetoothing to a phone.
Dude, Governments have the power to tax millions of people. They don’t need your piddling bribery money. It’s too much work for too little reward.
Law enforcement and intelligence agencies find this sort of information useful to further their agendas.
Oh wait, you're probably talking about the United States? It's scary how fucked up that place is just millimetres below the surface.
https://en.wikipedia.org/wiki/FBI%E2%80%93King_suicide_lette...
They didn't bother to get the servers running before pushing out a gigantic advertising campaign shaming anyone for not using it.
... Despite it having obvious flaws from day one, that showed it was mostly a cut 'n paste of Singapore's GPL app. (Though you can't access the source. National security trumps freedom of information and promises.)
Have you got something supporting this? Here's a panel of Australian-based security people decompiling and discussion the details of the app: https://www.youtube.com/watch?v=U3dN99ljgD4 Are you saying they've all publicly admitted to committing a crime and are unaware of those laws? Are all editors here https://docs.google.com/document/d/17GuApb1fG3Bn0_DVgDQgrtnd... criminals?
The only serious analysis I can find is in http://www.austlii.edu.au/au/journals/JlLawInfoSci/2003/2.ht... and it's "kinda depends why you're doing it, but either way it's largely untested".
> A person must not decrypt encrypted COVID app data that is stored on a mobile telecommunications device.
That video shows them looking into how the data bundle is assembled, but I don't believe they actually touch it or run it in an emulator, which would very much breach the determination - because unless you're one of the exceptions, you're not legally allowed to run the software outside of tracing.
Exceptions are given for those in employ of the health department, or other government bodies.
Whilst that might vaguely not mean decompiling the app, the minister's own press conference is clearer on the intent [1]:
> It cannot leave the country, it cannot be accessed by anybody other than a state public health official, it cannot be used for any purpose other than the provision of data for the purposes of finding people with whom you have been in close contact, and it is punishable by jail if there is a breach of that.
Decompiling the app steps outside the provisions for looking at the data, and yes, you don't have permission to look at your own data.
[0] https://www.legislation.gov.au/Details/F2020L00480/Html/Text
[1] https://www.health.gov.au/ministers/the-hon-greg-hunt-mp/med...
The determination is clearly aimed at data usage and prevents people from trying to decrypt the reports from other users. The whole fragment of the interview is about the data produced by the app and how it should be protected as sensitive information. I can't see anything there that would prevent you from reverse engineering "to see how any of the app is running."
It's not even obfuscated or protected from decompilation in any way, so it's trivial to look at with static analysis tools. (i.e. without trying to run it)
Even the headings don't mention the code: "Collection, use or disclosure of COVID app data", "Treatment of COVID app data", "Decrypting COVID app data", "Coercing the use of COVIDSafe".
Perhaps more than the intent, but this is a government that doesn't deserve the benefit of the doubt.
Circumventing any "access control technical protection measures" is currently a crime under Australian law (Section 116, Copyright Act). They may well consider any decompiling tools to fall under that particular law, as well as use of said tools.
In March, they pressured a university in firing someone researching into their own data breach to see how bad it is. [0] There isn't a law against de-identifying, especially when it is in the public interest, but they went ahead and threatened severe legal action anyway. Whilst simultaneously claiming that said data breach doesn't contain any personally identifiable information.
They had to be taken to the High Court to be shown that an algorithm cannot be used as evidence that a debt exists, and that decision makers actually need to do more than just trust the system. [1]
If it embarrasses them in any way, then they are not above twisting laws to suit them. [2]
[0] https://www.theguardian.com/australia-news/2020/mar/08/melbo...
[1] https://www.theguardian.com/australia-news/2019/nov/28/robod...
[2] https://www.abc.net.au/news/2020-02-28/abc-not-appealing-fed...
https://webcache.googleusercontent.com/search?q=cache:EJWgZa...
Reading through its text, I can't find any clause having the effect that you describe.