Since it seems that false positives are uncorrectable (given that even a New York Times article hasn't gotten this one fixed), and the consequences of one are thus life-destroying, even a single false positive should mean shutting the whole thing down.

Google is not unique here, nor does it have unique power. This story would be just as plausible, and difficult, if the company in question were Facebook or Apple.

Once a system exists to scan for this material, it’s quite easy to argue that shutting it off would be immoral and potentially criminal. “You had the means to identify abused children and help them, but you turned it off because you couldn’t work out a reasonable policy to deal with the false positives? Try harder, people.”

(Note that I do not speak for my employer here.)

What they should shut down, is their irrevocable deletion of accounts. When the false positives trigger nuclear bombs, then yes, a single false positive is too much.

As far as I know, only Google does this, so everyone else can keep scanning in this argument.

It's curious how the big providers have been scanning for CSAM for YEARS with nothing making the news...because hashes are much different and don't false positive like this.

Cut to just a few years later, and here we are: "shutting it off would be immoral and potentially criminal". We can't go back! So, that clicking sound you hear is the ratchet turning.

Obviously CSAM is vile and not something your average citizen condones. But so are those other crimes. I suppose children "pull on the heartstrings" more than your average victim, but children are also the victims of those other crimes. Why haven't we started surveilling personal messages to detect criminal planning in advance? Because more people understand the nasty implications of the government reading your messages, but few understand the implications of machine learning "detecting" CSAM in your Google Drive or Photos?

The planning of a crime can be illegal itself, eg “conspiracy to commit murder” is illegal, but generally it’s fraught with existing legal minefields.

See e.g. this open letter to Facebook authored by US AG Barr and counterparts from the UK and Australia. https://www.justice.gov/opa/pr/attorney-general-barr-signs-l...

I'm always amazed at the number of people who come to a complicated nuanced issue like this, accuse one side of having implemented an absolutist and inflexible dystopian monster, and then promptly demand an equally absolutist solution.

The libertarian answer might well be that we need to tolerate CSAM in the interests of free speech, but that's not the sense of the society we live in, nor of the laws it has passed. All paths must be middle paths.

One might argue there are products that can fill this niche and certainly many on Hacker News are capable of setting up their own home servers. However, that is not nearly as convenient or realistic for the average person to do. Those setups also don't integrate directly into default services like iMessage, Photos (including both storage and things like photo tagging), etc. Modern desktops and many people's internet connections should be powerful/fast enough to handle this.

To the extent user data is used to feed big-data algos/machine learning/etc, this might not be practical. That's why I suggest Apple do it rather than Google (Apple already claiming much of their ML services are handled on-device).

One might argue that's essentially something iTunes handled (still could?) but again it's not nearly as integrated as modern iCloud services. And that only works over a local network or direct wired connection.

The basic idea is just to move data hosting and processing from third-party servers where users have no direct control over how their data is used or accessed to hardware sitting in their home.

With all of the services you mentioned.

Then there are services like email, iMessage, reminders, notes, etc. The latter there may be some service that does it but not in the default apps. The former I don't know of any plug-and-play solutions not cloud hosted. For example, many privacy concerned people use Proton mail but that is cloud hosted and subject to search without the user's knowledge (hopefully provided a valid warrant, but I don't think that takes away from the point).

I'd also rather services like find-my-iphone, recent locations, and other map data be handled on a server I control without Apple having any potential access.

I'm not saying there are no solutions but I am saying that they involve tradeoffs that a first-party solution could make a lot smoother. I'd gladly pay for a home Apple server that could replace all iCloud functionality without having to change my habits. I suspect I'm not the only one - just look at their recent privacy focused offering that turns off a bunch of features for "at risk" individuals.

This is the hook that Tim Cook has been setting up for years. They wanted you to take a bite out of their services without thinking about how their concentration of power effects you. Now that people are realizing that iCloud/Apple Pay/App Store are all just ploys to add an Apple Tax to other segments of the economy, they're relying on the people addicted to their services to justify it's use. When people talk about the degrading Apple ecosystem, this is what we're referring to.

I jumped off that racket years ago. Buy a few WD Reds and set up RAID, then you can self-provision your own cloud resources.

That would be very cool but, sadly, 60% of that market is in this comments thread. And we want it to be open source.

I don't see how that bit would be a problem. The market size, yes, it's a problem, but it's supposed to be a cloud service, there is no loss on the clients being open source.

Users who aren't tech savvy who are trying to run their own mission-critical software on their own hardware:

1. Should expect to pay $200+/month for a support plan. That will buy them one hour of an engineer's time per month. They are going to need that time, because deploying software is hard, keeping software running is hard, and there are no SRE/DevOps economies of scale when every Tom, Dick, and Harry wants to run their own service on their own hardware.

2. Should expect to either deal with all the problems (stability and security) that come with regular auto-updates, or regularly getting pwned (because they aren't auto-updating, and because they are a juicy, high ROI target for bad actors).

3. All this stuff already exists for enterprises, and they aren't complaining in these HN threads, because their needs are being met by cloud providers (As are the needs of politicians. Because they generally don't need security, they need security theatre. As long as they can check the compliance checkbox, they are in the clear.) The people whose needs aren't met are the long-tail small-fry of journalists and hobbyists and their ilk. Who aren't enough of a market, and can't afford the money, and the time, because see points #1 and #2. Like, maybe the NYT can afford to have this set up, and set up well for their people, but they already have dedicated IT teams doing just that!

It's really a supply side issue pretending to be a demand one. Self hosted services of all kinds are woefully underdeveloped.

Tech giants aren't going to encourage you to own your data and hardware by spending effort on supporting it, because that undermines key reasons for building cloud services: surveillance and control.

On the contrary, it would make a lot of sense for Amazon to work on open source self-hosted server apps, because they would seem to be the answer to the question, "Why would a non-techy want a cloud VM." Surely their PMs must day-dream about a future in which a cloud VM is as common a monthly fee for a middle-class family as a gym membership or netflix subscription, right?

If the server software were confined to macs, that would also act as marketing just like iTunes was mac only for a while. Though I recognize Tim Cook seems far more interested in selling additional services than hardware unfortunately.

Unfortunately, iCloud works well for probably 99% of their user and, like someone in this thread said, the only people interested in this service are in this thread.

Many, many HN users want open source primarily for ideological reasons. A closed-source solution (or even a source-available one) is unacceptable for that reason alone, and not related to cost or practicality.

I think Synology is the closest product I know of that provides something like what you're describing, though, as you point out, it's not OSI open source.

I don't see a solution for this. Billions of people and probably a ton of people reading this on hacker news like me have their stuff at google or another top cloud provider. Google causes untold suffering for people by not let people recover their accounts. There was the recent librarian letter requesting help for people this happens too. We might know someone at google to ask for help. Wonder if this person did.

Most importantly, due to users hosting all the data using their own physical space, electricity, and bandwidth, the incentive for Synology to scan everything stored evaporates: they are not on the hook for illegal files hosted on a server in someone's home. I think the picture is less clear when it's Google or Apple doing the hosting, which leads them to scan people's files as a policy.

I don't know, this sounds like a technical solution to what is a political problem. Gmail, today, already scans your email for ads; what if congress decides that they started doing CSAM on your emails? How feasible is it for everyone to host their own email in that situation?

This also echoes the controversy with CSAM scanning in iCloud. There is no clear national consensus if the government should be allowed to invade your privacy when it comes to child abuse. Ironically, terrorists seem to have better privacy protections than a pedophile.

Personally I don't think CSAM scanning of your private data should exist, much like I think everyone should be afforded E2EE communication even if they might break the law with it. Whether we allow our privacy to be invaded in such a manner is a political/regulatory problem. A "private cloud" doesn't help if congress decides that comcast should also be allowed to MITM your SSL connection to scan every image you donwload.

https://support.google.com/mail/answer/6603

> The process of selecting and showing personalized ads in Gmail is fully automated. These ads are shown to you based on your online activity while you're signed into Google. We will not scan or read your Gmail messages to show you ads.

Here is an except from my "welcome to Gmail" message, from Google, from June 2004:

"You may also have noticed some text ads or related links to the right of this message. They're placed there in the same way that ads are placed alongside Google search results and, through our AdSense program, on content pages across the web. The matching of ads to content in your Gmail messages is performed entirely by computers; never by people. Because the ads and links are matched to information that is of interest to you, we hope you'll find them relevant and useful."

Not to be overly pedantic, but I don't think this is true on a technical level? Any "private cloud" worth its salt would presumably use its own E2E encryption, with keys known only to the owner.

The government could force the private cloud vendor to build a back door, but that's kicking things up a notch. The vendor could also decide to build a back door on their own, I suppose... which is another reason people in this market tend to want open source.

It also doesn't help if you send the image in an email to a third party that doesn't share encryption keys with you. Which is going to be the case for situations like the one described in this article unless your "private cloud" expands to include all your health care providers--not to mention your bank, your insurance company, etc., etc. Which of course it won't; all of those third parties are not going to care what your private cloud setup is; they're going to use whatever all the other large companies use.

Absolutely. But who's going to sell it to the masses? Most people don't care, as they haven't been told to - no one has sold it to them.

And who is going to explain about the erosion of privacy? Would that be Google, the government? Why, when both are beneficiaries of the existing system (ie making money for google, increasing control and monitoring for govts)? This is a market that is studiously ignored as it is in no one's interests!

Do you remember when Google piled into rss only to try its best to then kill it? When data is your business, it makes no sense to cut yourself out of the intermediation.

NAS vendors saw that future and worked toward it. Synology's solution is pretty extensive and could have been a real game changer... if it wasn't for:

- being a PITA to handle network shares on the mac

- iTunes being the prevalent link to iPods and iPhones for such a long time, even for local file exchange

- music files being the sanctuary no third party messes with on iPhones

- photo sync being such a PITA as well for anything non Apple for so long (still is?)

- iOS backups being iCloud only

- even Time Machine backups require more specific adjustments that make it harder for normal people to use outside of the Apple offering.

Don't get me wrong Google also screws the market in so many nasty ways, but Smartphone got rebooted by Apple, and they royally fucked it for so many third party applications to cement their own services' profits.

This is the only way I am aware of to maintain a life with most modern technical conveniences and still have the protections or the Fourth and Fourteenth Amendment.

If anyone wants to pursue this type of digital sovereignty, I founded an organization called #! that will happily teach anyone interested. https://hashbang.sh or #!:matrix.org

Joining the matrix room is also a valid option.

Given the reach, it's absolutely time for "government interference." Now, I'm not actually saying that we definitely should shut things down. But we absolutely need to drag Google people into public hearings and such. Google et al have inserted themselves into their lives to an extent that they ought no longer be able to claim "but we're a private company." It's increasingly difficult, if not impossible, to opt out.

Example; I fully intended for my children to not have anything like a Google account until a certain age -- but then the pandemic and remote schooling happened. It is true that with a great deal of pain to teachers et al I could have opted out, but that would have been literally unreasonable, not to mention unsustainable for other people.

It is easy to ask for public hearings but if there’s no intended remedy or change they will just be a show trial and nothing will change.

Your example is great, because to me it illustrates my point well. What would you like to have had be different?

Warning labels today, I think means "fixing voluminous EULA crap," which we've done in other fields reasonably well, e.g. the FDA I think is pretty solid here. Simply requiring "readable" terms of services I think would go a long way (though you have to wait a bit for things to shake out.)

The general principles behind "common carrier" type services would also be a good idea to start thinking about; they essentially just say "When your product or service becomes a day-to-day thing for a lot of people, we're going to hold you to certain accessibility rules."

As I see it the fundamental problem is not that Google isn't behaving as a responsible steward of free speech. The problem is that Google shouldn't have this power to begin with. By law, Google should be allowed to capriciously and arbitrarily shut down a user's account. It's their service, they shouldn't need to provide a justification. But Google shouldn't be in a position where they control so much of the market. There needs to be real competition in this area so users can easily switch to a different provider if they lose faith in Google's reliability. Where the government needs to be involved is in ensuring real choice in a free market, rather than the current duopoly.

Ensuring a free market isn't easy. It's much easier to simply pass a law that says that Google (or a company in a similar position) needs to provide better service or offer users recourse to dispute this sort of service interruption. But that's just putting wallpaper over a crack. We need the government to commit to the hard work of enforcing open standards of interoperability so smaller entrants can get into the market.

Just pass the regulations and competition will likely find a way around it. Honestly, this to me is the biggest blind spot in all of "pro free market capitalism." The good flavor of capitalism is antifragile, You can make competition better precisely by making things HARDER on private companies, not easier. Stuff gets better because the weak die.

If we've allowed Google (and a few other select tech companies) the privilege of becoming de-facto monopolies in their space (and I believe we absolutely have, often intentionally as government policy in the international space) then they should be bound by the same rules that the government is.

I think this is even more obvious when you consider how some of these companies creep into government processes.

Ex: Google classroom is estimated to be in use in more than 60,000 schools across the US. This means that a Google account is required for all students in these schools (opt-outs exist, but are impractical, hard to enforce, and often cause you to be labeled as troublesome by teachers and staff).

If my child attends a public school that is using Google software and requiring a Google account - The government damn well is involved, and I expect the government to require the same rights & due process for a Google account as they do for other governmental actions.

I strongly disagree.

While there should be some level of protection for large companies ability to terminate customers at will, there also need to be consumer protections against unreasonable EULAs and capricious decisions. (Edit: some such protections already exist, so it is a question of if and how we shift the balance between consumer rights and corporate rights. I think as the companies grow larger, more powerful, and more integral to our daily lives, that balance needs to shift further towards consumer rights given the power imbalance.)

When companies sell us digital goods and become integral parts of our digital lives, there arises a need to balance their rights of free association against the rights of their users to due process.

That is not to say that simply mandating a fair dispute resolution system is a sufficient solution, but it is a necessary step.

I think mandating data portability and maybe even interoperability is also necessary (though an interoperability mandate seems tricky to do well.)

Google is not selling you any digital goods. They provide services like GMail for free to people. That is a big part of the problem: users are not paying customers and so have no leverage, as they would in a proper customer relationship. Instead, users are the product, and the actual paying customers have all the leverage in determining how the product is treated.

But that being said, I meant the "and" following the statement quoted to be an "or".

The cliche of "if you aren't the customer, then you're the product" leaves out the important detail that you can always be both.

Being a paying customer is no protection when you are in a business relationship where your have no power, especially when tou have few alternatives.

None of which are relevant to the issues of privacy being discussed. The relevant Google services, such as GMail, are provided for free (meaning no money is paid); that's why so many individuals and institutions, including government-run schools, shortsightedly adopt them without considering the actual (non-monetary) costs.

> Being a paying customer is no protection when

...you are only a paying customer for products that have nothing to do with what you're complaining about. Google isn't going to care any more about your GMail privacy just because you happen to have bought some books, movies, music, or apps on Google Play or own a Pixel phone. They have demonstrated that quite clearly for years now.

The issue being discussed is a company like Google banning your account with no recourse. Privacy only comes into play in this discussion because the given reason for the ban comes from the scanning of Google Drive uploads for CSAM, which has nothing to do with Google selling your data or targeting ads at you and is done regardless of whether you pay for Google Drive storage or not.

Google operates a single account across all their services, so a ban for a Google Drive CSAM false positive means that you lose access to not only the contents of your drive, but also any books, movies, music, apps, cell service, advertising campaign, hosting, etc.

We have dragged Google into public hearings. The problem is that the purpose of these hearings has always been to collect soundbytes rather than to try to define a relationship between internet services, the people, and the government.

No, it's where we the people need to understand that rights come with responsibilities. If you want to claim the right to keep your data private, you need to not give it to third parties that you know are not going to honor that. And knowing that is easy when it comes to Google, Facebook, Twitter, etc., because of the simple rule that if you're not paying for the product, you are the product.

The problem here is not too much free market and libertarianism, but too little. The ultimate reason for this...

> It's increasingly difficult, if not impossible, to opt out.

...is that the government has inserted itself more and more deeply into our lives, and we have let it. For example, your kids' schools are beholden to Google because they are run by the government, which is insulated from free market competition and can just punt on providing proper infrastructure. And the reason why Google, Facebook, Twitter, etc. can survive on the ad-supported business model, instead of having to make their users actual paying customers as would be the case in a real free market, is that the government has set things up to favor them. They can get cheap loans to build the massive infrastructure they need because of government monetary policy. They can insulate themselves from any effective legal challenge because of the way the legal system is set up.

Private schools are just as beholden to Google. No school is going to compete on the basis of "we don't use Google services (we pay for more-private infrastructure)", in any world, ever.

> we the people need to understand that rights come with responsibilities. If you want to claim the right to keep your data private, you need to not give it to third parties that you know are not going to honor that.

So if I'm a parent of a school-age child, what exactly do you think I should do? If I don't decide that Google accounts are enough of a reason to homeschool my child, will you claim I don't understand that rights come with responsibilities, and that I (and my child) don't have any right to expect our data to be private?

We live in a society. Sometimes the only way to crack collective action problems is government intervention.

Um, what? Didn't you just get through saying you don't want your kids to have to have Google accounts just to go to school? Do you think you're the only parent who wants that? Isn't the protection of your kids' privacy by the school worth something?

If you're unwilling to pay anything more than zero for something you claim you value, then your claim that you value it becomes, to say the least, questionable.

> if I'm a parent of a school-age child, what exactly do you think I should do?

You mentioned homeschooling yourself. In our current environment, unless you are very lucky in what school district you happen to live in, that is probably the only option to avoid giving up your kids' privacy.

> don't have any right to expect our data to be private

Google apparently doesn't think so, and neither do your kids' current schools (and I suspect your implication that the vast majority of schools in the country are similar is correct), which means the government that runs those schools doesn't think so either.

> Sometimes the only way to crack collective action problems is government intervention.

No, that's my point: governments do not solve collective action problems. They grab power using the excuse that they are going to solve such problems Real Soon Now, but they never actually do. Governments run all the public schools in this country: they have had all the tools they need to solve the collective action problem of how to protect kids' privacy in a digital world for quite some time now. And they don't care. And this is just one of many, many examples.

I'm willing to pay something for it. That doesn't imply I'm willing to pay an unlimited amount for it. Google accounts are not in the top 10 things I care about with regard to schools.

> You mentioned homeschooling yourself. In our current environment, unless you are very lucky in what school district you happen to live in, that is probably the only option to avoid giving up your kids' privacy.

My mention of homeschooling was intended to drive home the point that the only way to avoid Google in this situation has incredibly high cost, out of proportion to its benefit. If you truly think that a parent should homeschool to avoid giving data to Google, that's an extreme view.

That does not mean I don't think giving data to Google is harmful!

> No, that's my point: governments do not solve collective action problems. They grab power using the excuse that they are going to solve such problems Real Soon Now, but they never actually do.

Apparently your solution to collective action problems is that individuals must take action, without coordination, even when it's individually counterproductive. And you view failure to do so as indicative that they place no value on the problem getting solved. In other words, you seem to dismiss the notion of the "collective action problem" as a phenomenon at all. Frankly, I find this view absurdly naïve.

No, I'm just objecting to the claim that governments can solve such problems.

> the only way to avoid Google in this situation has incredibly high cost, out of proportion to its benefit

That's because the current situation was created by a collective action failure, one which governments have not done anything to help and have in fact made worse, by doing the same thing individuals have been doing: shortsightedly taking advantage of "free" services from Google without considering the actual (non-monetary) costs.

Governments are not going to fix this situation. The only way to fix it is to invert the collective action that caused it in the first place. I agree that, because of the depth of the hole we have dug ourselves into, climbing out of it is a huge effort. That doesn't change the fact that the problem won't be solved any other way.

Clearly Google already has a process to escalate detection to the police so banning the account based on what happens there doesn’t seem like a big leap.

The police aren't going to recommend a ban. They don't want people banned. They want people arrested, tried, and convicted for criminal possession. They're going to recommend keeping the account open until they have enough evidence to take it to court.

Google probably doesn't share the same rosy outlook. Long-running investigations that do not result in charges will be bad PR. The headline "Google shared private data on 100s of innocent users with law enforcement" will not be received well.

It might be helpful to frame any cloud-based scanning of user activity or files as akin to mass-testing in health. While it sounds good in principle, there is always a background rate of false positives. And being incorrectly diagnosed can have negative consequences. There has been a large amount of examination of the consequences of mass-testing for various diseases without indication.

e.g.

COVID testing: https://www.bmj.com/content/373/bmj.n1411/

Herpes: https://www.statnews.com/2017/01/26/flawed-herpes-testing-le...

Genetic testing: https://www.geneticsdigest.com/genetic-testing-potential-con...

And the law is about to go much further:

https://www.wired.com/story/europe-csam-scanning-law-chat-en...

Note that these laws don't mention platforms by name. So not copying your chats and images to the authorities, even going so far as using an application you custom write for a particular conversation, from even social workers and certain (not even named) private organisations (not even just the police or justice departments) will be illegal regardless of the content. They haven't even bothered to list which institutions are to have access to your chats. Just "law enforcement". Do I need to copy Latvian law enforcement on a message I send from France to my wife in New York? It's actually not clear.

Any point where the law would require a little bit of effort by government, everything is left open. And these laws require a LOT of effort and infrastructure by governments to even enable citizens to comply with them.

Call me a cynic, but I find it very hard to believe EU or US citizens would vote for their private chats or pictures to be copied even to the police, or scanned for anything at all.

Never mind that, of course, said authorities haven't even bothered to make it possible for other platforms to do said reporting (I guess Facebook, Google and Amazon just find phone numbers to call somewhere, and then send unencrypted email with the material, because anything else just wouldn't work). Of course there is zero mention of what a person can do if such reporting mechanisms are used to discredit them (for example, send CSAM to your ex/political adversary, call the police on them).

And of course, the police, social workers' and other institutions reputations where it comes to helping victims of CSAM is beyond terrible. In a lot of countries, including the US, cases where people protect children against CPS, even in cases of abuse, are easy to find. The reason is the extremely bad treatment child victims receive once reported, in other words: these reports aren't helping the victims at all, and this is a problem well-known for decades. It's not likely to improve any time soon. Cases of children fighting against these institutions when they are the victim aren't just well known ... they're actually the norm, the expected outcome of such interventions.

I'm against conspiracy theories ... but I do agree with them on one point: I find it very hard to believe laws like this are even meant to protect anyone, other than the institutions stuff is being reported to themselves. If they were meant to protect, then we would make sure the treatment of children after reporting would be great, AND THEN go look for children in problem situations. The opposite is happening. You want my guess? These extreme measures are attempts to minimise the times government institutions get embarrassed by (the extremely rare cases of) serious crimes against children, which are still mostly missed.

> Mark asked if Mr. Hillard could tell Google that he was innocent so he could get his account back. “You have to talk to Google,” Mr. Hillard said, according to Mark. “There’s nothing I can do.” Mark appealed his case to Google again, providing the police report, but to no avail…

> A Google spokeswoman said the company stands by its decisions, even though law enforcement cleared the two men

Second, what is Google to do? Clearly the picture is not ok, I mean hopefully we can agree on that. Yes it isn't reason to suspect child abuse, but it isn't ok to spread either, which is of course what Google is worried about.

Going through the account and removing this picture is ... also very much not ok. Hell, not even just because of the privacy invasion that would be. On what basis would they declare things safe? There is no guidance in the law on what is allowed.

The government just throws the entire thorny problem on Google and then declares itself not available for comment.

So I don't see what you want Google to do here. I don't see what options they have. Restoring access to the account may be a crime for Google (that might turn out to be "knowingly" allowing the spread of CSAM). Going through the account and removing the material is fraught with problems too, starting with the impossibility for Google to declare material safe, as well as privacy issues.

[1] https://en.wikipedia.org/wiki/Corporate_personhood

Yes, Free Software culture has always been more about justice on a new (digital) frontier where new forms of injustice and abuse roam wild (and have only grown worse).

I sincerely think the best model for understanding big-tech corporations like Google is as serfdom under feudal warlords within modern fiefdoms. It closely mirrors these historical power relations where laws and constitutions have nothing to say.

The article is long and complex but I eventually found the kernel in this line:

> "A Google spokeswoman said the company stands by its decisions, even though law enforcement cleared the two men."

Though the moral questions behind it all are very complex, this case is not itself actually that complex. There was no mistake. No lack of proper investigation or tardiness by the police. The child, parent, doctor and police - all of the parties except Google - acted fairly, in good faith, mutuality and consent.

Google is the problem here, and simply believes itself a law unto itself, that's the nub of it. For all the posturing about complying with the laws of nation states, companies like Facebook and Google have grown smug about their power. When they roll over the toes of the innocent, they laugh and say "and what are you going to do about it?"

We are in new "might is right" times and nobody big enough has yet had the courage to say "Act justly, or we will hurt you back", and then follow up.

For the ordinary citizen, the only sensible course is to resolutely refuse to use their services and products, and Free Software which provides so many alternatives to Google, Microsoft, Facebook and suchlike is the solution. It is the moral choice that a good citizen can and should make in these times.

the most frustrating question I have - Why isn't Google restoring his account despite him being cleared by the police ? Why are they not even explaining this ?

As the last paragraph of this essay points out: tough. Google needs to step up and deal with the consequences of its own size.

This is where regulation can help.

One case could be this one: doing the right thing could open you up to attacks from bad actors. A regulation could give you air cover (and legal cover).

Or lets say you want to be more environmentally friendly but customers won't pay extra. Your competitors do too, but of course you can't agree to do this (that's collusion). A regulation removes the risk and is applied to all.

It's one thing to get the attention of some Googlers on the front page of HN, it's a whole other order of magnitude for the front page of the NYT.

I cannot imagine how this hasn't become an urgent priority for Sundar himself to make sure this gets fixed and fast, at a minimum in these known situations. What could they possibly still be debating or deciding inside there at Google...? Police determined no charges, so restore the accounts. I don't see any potential risk, legal/reputational/otherwise, this would open up. Google's inaction here boggles the mind.

[1] https://www.nytimes.com/issue/todayspaper/2022/08/22/todays-...

When I was at Google, I was continuously amazed at how hard it was to stop this kind of own-goal when we knew it was happening. (Based on where I was inside Google, I came across (dramatically-less-sensational) obviously dumb decisions of this sort).

It was always shockingly hard as a Googler to get traction on fixing some customer's bureaucracy-navigation problem -- but usually it getting into the press finally got someone through the bureaucracy.

Rather than improve their AI to avoid more of these types of false positives, or invest in their customer support teams to deal with these situations, they can eliminate this entire class of false positives by scaring people with articles like this.

EDIT: which if true, is kind of extra fucked up when you consider the potential consequences of delaying medical treatment for a child.

We saw similar stuff when banks were getting bailed out because they were "too big to fail". That meant that their very size was a threat to the financial system - they were really too big to allow to exist.

Think about Google's alternatives here.

They could (1) not have CSAM filters -- your data is private no need for Google to scan anything -- in which case people would use their platform to distribute illegal content and they'd be a nice target to get massive damages from.

Or (2) give people due process, rights, appeals, etc despite not having any of the tools of a court of law, it's even illegal for them to look at the evidence, and if they make the wrong decision or they make the right decision too late they're still liable for their platform being used for illegal activity. Keep in mind the only way to truly know if material is sexually explicit is to show it to Potter Stewart. He'll know it when he sees it.

Or (3) be proactive in building systems that detect illegal use of their platform and aggressively, without process or appeal, remove any user of the platform the moment you have any evidence at all they are engaged in this activity.

Because people will try to use Google's platform for illegal activity and Google knows this, (3) is not only the only sensible option but it's actually the only legal option. It would, in effect, be illegal for Google to do anything other than scanning all your files and aggressively ban people.

I'm not a fan of this phrasing because it absolves Google of responsibility. If the CSAM is on a Google server then Google is possessing and distributing CSAM. You can say they're doing it unwittingly or unintentionally but that doesn't change the fact that they're doing it.

Amazon isn't scanning anything, where are the damages paid out to those using AWS for CSAM?

They don't need any of the "tools of a court of law" you mentioned in (2) if they defer to the authorities and then stay in their lane.

It seems to me that the big issue here is Google digging its heals in on an account ban, even when the victim of a false positive had been cleared by the police.

Aren't blackrock/vanguard making the laws in the US? Like killing the planned obsolescence related bills?

>BlackRock’s spokesperson said the votes can also be carried out by a portfolio manager – and in some cases at BlackRock, can be carried out by the clients themselves (here).

>BlackRock and Vanguard do not “own” all the biggest corporations in the world. They invest trillions of dollars into leading companies on behalf of their clients, who ultimately own the shares.

https://www.reuters.com/article/factcheck-business-investmen...

That said, it is actually blackrock/vanguard ppl on the boards, and often they are the biggest ones, and expecting no steering from them or the "big" investors behind would be... "unexpected".

Their is also the massive debts with the gigantic interests to be paid all along the calendar year. In the case of starbucks, owners of the debts is public info or the web page is incomplete?

That said, now I understand why online non-franchised/licensed starbucks stuff has the aspx file extension.

Citation needed.

I think error-prone systems are still necessary today to reduce costs to a tolerable modern level. I'm not inclined to throw out the baby with the bathwater regarding communication costs for the same reason I don't want $300 toasters even if "they don't make 'em like they used to".

A good example is something that Facebook does: they have a blacklist of domains you can't send on Messenger, the message will be instantly blocked when you paste a detected link. From experience, this list is fairly large and very inaccurate, but the false positives for this system are annoying rather than catastrophic because all it does is block a specific message. If you don't like this, then you can use Signal or something, I don't want other people making that choice for me.

Suppose Google analyzed all business data that passed through its servers looking for patterns of tax evasion. If the probability of tax evasion was high, reporting it to the IRS, which offers rewards, could be a profit center.

[1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2394843

https://youtu.be/PdOx36bZMBY

Good recent is example is Facebook suppressing stories that the FBI wants suppressed.

That book is very relevant today.

[1] https://quoteinvestigator.com/2017/07/28/flag/

Instead, Facebook should have been promoting stories that the FSB wanted promoted. :)

What will USA 2.0 look like? Is there a way to work on this without it being a bloody revolution? If we got 2/3rds of the population to green light a refactor of the constitution could it happen? Who would do it? Who are the right people?

Importantly, if we want to preserve the spirit of the USA's founding documents, they need to evolve more formally and officially than supreme court case law and bloated regulatory agencies tied to deeply industry integrated lobbyists. I really wish writing laws was more like the RFC process.

I don't think the solution is to "just run more stuff at home" as much as that tickles my technocratic jimmies.

Obviously such a system would be way more complicated than I can describe in a single HN comment, but I do think we have the technology to build a much more public, accessible, and directly democratic method of writing laws.

As an aside, did you know no one even knows how many federal laws exist?[1].

[1] https://blogs.loc.gov/law/2013/03/frequent-reference-questio...

I could even imagine a voting system that lets you name your favorite politician as your proxy. So politicians could collect the votes of their base into a block. Maybe even a mechanism of saying I want this expert to cast my vote on x topic, and this other thought leader on y topic. Independents might retain their vote on every issue if they care to.

I'd probably go with a system with tiers of votes as an issue gains more public attention. Say a minor change is proposed on a niche law and those who are passionate about fishing rights in the Colorado River get involved and vote. Say 5k people vote. Then there's a cooling off period before implementation but all of a sudden the issue ends up in the news and 5m people are now interested so there's another vote with another cooling off period. Only this time 10m people would need to express interest to hold another vote.

I don't know the perfect system but I'm confident we can come up with a more democratic process than what was possible 250 years ago. One in which the public more directly decides what issues get voted on.

Congress has limited capacity and can't focus on everything but the work can be distributed if it were more open. A system in which experts could make proposals directly instead of being forced to educate and convince a politician the issue is even worth spending political capital on. And a system in which citizens can see all the changes being made publicly. It seems every time a major bill is passed it tends to be this huge amalgamation of new laws that often aren't even related. And then news articles follow pointing out how x, y, and z things got slipped into the 800 page bill no one in Congress actually bothered to read.

And a single official repository of all laws that can be edited directly and kept up to date. The United States Code is actually a pretty good place to start, here I’m just suggesting modernizing it.

There's nothing worse or more myopic than this kind of "pragmatic", numerical approach towards liberty. As if living in a panopticon cage is desirable, as long as the wardens never make a mistake.

Why is it myopic? Well, if being forced to use technology that betrays its user is okay as long as one number is sufficiently larger than another, then why not force all new houses to come with indoor microphones and cameras, always listening and watching - only for the most heinous crimes, of course!

Are you "unilaterally opposed" to such a measure, despite how much sexual abuse and domestic violence and murder it would prevent? What about tracking and shock collars for everyone? How much violence could be stopped if everyone was forced to wear those!

Don't like it? Not even if the false positive rate was zero?

If Google deputized a VP to hunt you down and murder you in your sleep, that also would not violate HIPAA.