Can we just ponder for a moment that we are being asked to believe that anything Twitter does or does not do could pose an existential threat to the United States?
“National security” is a sham now and has been a sham every other time it has ever been used as a blanket override for human rights. Human rights to free expression are more important than any nation.
This is just government spies trying to cover their ass whilst they continue breaking the law. They've conveniently mixed it in with "but terrorism!" to ensure that they get to continue breaking the law.
Hypothetically, a terrorist group could use private DMs on Twitter to coordinate an attack. If the US captured a low-level terrorist that didn’t know all the details of a plan but knew that the remaining details had been sent to other Twitter users, it would be a legitimate national security request if the government asked Twitter for copies of a user’s private DMs.
The hypothetical situation above would be an invasion of someone’s privacy, and any smart terrorist group should be using a message service that is end-to-end encrypted (I’m not sure if Twitter DMs are), but let’s acknowledge that there is at least a plausible national security reason for the US government to request things from Twitter. These requests can be abused, but any social network that has a private messaging feature could be used to coordinate national security threats.
Edit for clarification: If the US government made one request to stop a terrorist plot that they had some specific knowledge of, that would be okay in my eyes. If the US government requested access to every Twitter user’s DMs to scan for possible threats, that’s not okay. It seems like Twitter is trying to disclose the number of requests so we can see for ourselves.
> If the US government made one request to stop a terrorist plot that they had some specific knowledge of, that would be okay in my eyes.
How can we tell that they just didn't straight up lie about whether the one request was to stop a terrorist plot, and not to gather intel on political foes or ex girlfriends? With no oversight and no one allowed to advocate for the rights of the individuals under surveillance, I am still not okay with one request to which they claim is stop a terrorist plot that they claim to have specific knowledge of, because there's no checks and balances to prevent them from simply lying.
I cannot envision any terrorist attack that poses an existential threat to the United States. Fighting terrorism is not a national security matter. Even the largest terrorist attack in the history of the world (not counting the atomic bombings of Japan, which were terroristic by definition but generally excluded from that classification as being carried out during a declared war by a with-a-return-address nation) was at least an order of magnitude off from posing a threat to the national security of the US.
"National security" is one of those thought-ending phrases that's thrown about to make people stop thinking critically about the matter. Terrorism is extremely bad, but it's not an existential risk, by its very definition.
> let’s acknowledge that there is at least a plausible national security reason for the US government to request things from Twitter
There is already a system that permits the government, investigating crimes, to compel information disclosure from a provider. It's called a search warrant, and requires a (very cursory, heh) review by a judge.
The government is asking not only for the power to skip that step (they've already taken that - that ship has sailed), but to prevent people from disclosing just how many times that's happened.
"National security" is not an umbrella term that means "bad things happening to USians", although I'm sure that the people using it in this instance as a cover story for widespread, illegal, extrajudicial espionage activity by the government would prefer that you think of it that way.
Remember: every legitimate NSL could have been a warrant. There's already a legal framework for how to investigate and prevent crimes, including mass murder. Additionally, anyone who's ever read a US federal warrant for anything involving sequences of bytes or information technology of any kind knows that the burden is already hilariously bad/low, and that they're very close to being rubber-stamp already. I personally have never seen a federal warrant (they're written under penalty of perjury) that didn't have naked factual errors in it, and I've only seen the ones that got approved/signed by a judge. There's a 100%, easily falsifiable, perjury rate in them in my experience. AFAIK none of the federal agents who have signed these sworn statements have ever faced any negative repercussions whatsoever for the falsehoods contained therein.
So, that hilariously low/token bar to a search warrant is what they're skipping in these cases. NSLs and related FISA warrants are extremely suspect. It makes sense that they would want to hide and obfuscate their use and existence. They're basically carte blanche for the government to inspect any communications they want - no justification required. The only reason they'd use them in place of a search warrant is because they are illegitimate.
This is absolutely not about terrorism, or threats. There's already a functioning-for-them (albeit broken in the legal checks-and-balances sense) system for that in the form of search and seizure warrants; they're trying to confuse you with the terrorism card.
This is a cover-up for an illegal power grab, nothing more.
I thought a clear signal of something lacking E2E encryption is if a web browser can display messages without a client-only password. That is, if the company can render your messages in plaintext server-side, they can read them server-side and do anything with the plaintext.
Now I realize the client-side password could be saved and retrieved in the browser for ease of use, but I don't think that's going on.
Let's clarify Twitter request first: Twitter requested permission to reveal exact number of requests.
Companies are allowed to reveal gagged legal process in ranges of 1000, starting at 0, for six-month periods. Combined FISC and NSL requests can be revealed in ranges of 250.
This just shows that Twitter is fighting all the way, as I mentioned in my other comment.
The fact that Twitter refuses to enforce its TOS when the POTUS issues a call to armed insurrection against state governors is an existential threat to the United Ststes.
call your congress critters and tell them that they need to legislate the executive branch spinning up their own activitypub-compliant social implementations on government-funded infrastructure rather than continuing to relay messages through third-party commercial services.
What we’re being asked to believe is that the government doesn’t want you to see their backdoor search queries on Twitter’s database. That I do believe, regardless of whether there’s probable cause to think national threats among Twitter users exist. And hey, maybe it’s true... exposing government surveillance overreach probably is considered a threat by the people doing the surveillance, right?
OTOH, there probably are at least a few investigations into real cases of abusive or criminal activity. This may be getting used to watch what’s happening in other countries. There are a lot of possibilities for gray area questions that are neither clearly legal nor clearly illegal, and Uncle Sam knows that exposure could cause things to lean toward clearly illegal.
Let’s not pretend that privacy of Twitter messages is a human right, though. It’s a for-profit business that was already monitoring your traffic, and the government is asking for what Twitter already has access to. While the government may have a very different agenda, using Twitter doesn’t come with any privacy guarantees in the first place. Twitter’s not going to come arrest you, but they were never under a legal obligation to help protect you from the government or anyone else, and Twitter DMs are not private in a legal sense.
"We are empowered to stop terrorism." (or drugs, CP, etc)
"X can be used for terrorism."
"Therefore we are empowered to stop X."
Thus that agency gains power over all possible X. Like cash, any monetary transaction, photography, encryption, any messaging system whatsoever, or any number of other things that should make up the bedrock of a free society.
> Can we just ponder for a moment that we are being asked to believe that anything Twitter does or does not do could pose an existential threat to the United States?
Are we? The article just mentioned grave or imminent harm to national security, not an existential threat.
What is national security other than the continued existence of a nation? Any other sub-existential attacks can accurately be classified as "crime", which is a police matter, for which we already have a legal framework. They have to inflate it into something big and scary otherwise people would say "hey, wait a minute, that's not constitutional..."
Don't get it twisted: it's just a cover for domestic criminals inside the US government who have become accustomed to getting to break the law as much as they want for the last two decades. By making it seem special or important (it's not) they get support in suspending basic human rights such as publishing.
They can always pseudonymize and redact so that specifics are not revealed.
They have a point about a security risk but what they also sneak under that guise is their desire to hide intelligence capabilities and potential threats. It is not a risk if you don't have a specific threat or plan of attack,certainly not a risk to the nation. Vulnerabilty is not risk.
Revealing a request may thwart their investigation but the burden of proof that a specific threat that cannot be stopped if the information is public is required of them. Short of that, you can chit chat on the subway with a third cousin of a terrorist or something and they can say "if we can't monitor his DMs then we can't know if they are using code words to plan a terrorist attack"
On a different note, I think it is a failure if the US constitution to not clarify details like this. A constitutional convention to review and update the bill of rights and a few other things is long over due.
Yet an FBI agent can literally encourage the first ISIS attack in the US and it’s covered on 60 Minutes but no one bats an eye. We’re left just to assume this is the only time it’s happened. While the 3rd terrorist arrested months later assumed the FBI agent died in the attack — which is how involved he was.
Is the vague concept of national security that you raise the legal foundation for the actions the government is taking?
(of course the answer is no, it is not)
Sliding into Twitter DMs and baiting fools to say they are going to use fake explosives aren't particularly productive things for the government to be doing, but there's probably a reasonable legal foundation for most of it.
we are being asked to believe that anything Twitter does or does not do could pose an existential threat to the United States?
Well, they can’t have it both ways. Twitter takes credit for the Arab Spring which was indeed an existential threat to several countries, at least their regimes, but still millions of ordinary citizens were impacted.
What's more disturbing and I simply can't get my head around is that Americans feel that they can trust big corporations more than their own government.
I don't trust US corporations at all and I thought that there was some kind of corporate conspiracy to discredit the US government to allow corporations to gain more power...
Now I'm thinking that the situation in the US is even worse than that; not only are corporations trying to discredit the government, but the government is working hard to discredit itself as well.
There aren't many/any instances of single corporations causing >10M deaths, or oppressing millions of people.
There are lots of recorded occasions of single governments doing this.
Governments are seen as having authorities that corporations do not, and thus are to be viewed much more skeptically as a general policy. Companies can't raid your house with machine guns, or toss you in jail for publishing an integer on your website.
It's not that corporations are trusted more; it's that the maximum possible damage by corporate evil is much, much, much, much less than the maximum possible damage by government evil. IBM participated in the holocaust, but they didn't directly cause it. That needed governmental authority. It's prudent to maintain several orders of magnitude higher suspicion and/or skepticism of government activities; corporations are basically sandboxed (e.g. no armies, for the most part) in this model.
Corporate power comes, largely, through government power. Corporations coopt the government and use its power to protect or enrich themselves. Why would corporations want to weaken that?
> What's more disturbing and I simply can't get my head around is that Americans feel that they can trust big corporations more than their own government.
Government and corporations are merging to become the same thing. I just assume any data I give to a corporation or government is public since they share your data anyway.
> Now I'm thinking that the situation in the US is even worse than that; not only are corporations trying to discredit the government, but the government is working hard to discredit itself as well.
Of course it is.
Powerful and motivated interests have long been working to overturn basically every social program and every protection for labor, discrimination, environment, voting, etc. The goal is to "starve the beast," privatize everything to work toward a government that only runs the military and police, and to make sure it stays that way once it gets there.
Insidiously, they can win support by pushing the government to be more wasteful, more corrupt, less effective, more intrusive, more bumbling, and more frustrating to work with. The moment of victory is when they convince you that government fundamentally is a bad idea that cannot work well, all those taxes you are paying are just going to be squandered, and we'd be better off getting rid of the whole thing.
They have been at work for decades through groups like the Heritage Foundation, CATO institute, etc., to push this into the mainstream. They reached new heights of legitimacy with FOX news & co., and have made great progress with the Trump presidency so far.
We learn from Snowden leaks that Twitter was the only major platform that refused to participate voluntarily with PRISM surveillance. Twitter only complied with legitimate FISA requests they were not able to refuse.
Microsoft,Yahoo, Google, AOL, Skype and Apple all did more than legally required.
Not true. Twitter declined to build tools, but didn't withold information more than others. And since Twitter has no internal security, NSA could simply help itself to Twitter data without consent, whereas they might need to compel a handover from other orgs.
Everyone else you mentioned runs a non trivial business that interacts with the government significantly.
I know trump get's a lot of action on twitter but the next guy won't be using it for the main means of comms.
It's great that twitter took a stand, but if that company failed everyone would easily replace it. Twitter is toy.
Well I guess it depends on what your definition of "toy" is. I think all popular social media platforms deserve careful consideration. They have a colossal impact.
It's more interesting to consider what, exactly, government surveillance can be used for.
And really, "James Bond" style villainous plots aren't what we _really_ need to worry about. The real weapon in these platforms is disinformation campaigns-- and the actual actors behind such campaigns (as well as anyone else who is paying money to influence any user) should be revealed to the public.
Does anyone know why judges are so deferential to the US security apparatus? It seems in general like they have a clear bias toward "taking the security system at its word", even when common sense would dictate otherwise in any particular case. It's sincerely puzzling to me why national security entities seem to get treated so differently and I wonder if there's some legal precedent or consideration I'm not aware of.
Like in this case, I fail to understand how releasing the number of requests would have any security bearing.
If what you mean is why do most of the judges tend to rule the same way in this area, it's due to precedent, they don't really have a choice. The balance (or lack of one) you note was mostly established by higher courts years ago.
No judge wants to be responsible for a future terrorist attack or degrading intelligence capabilities against other foreign adversaries, regardless of whether the threats are real, perceived, and/or imagined. In this case, releasing the numbers might shine light into their capacity for investigations. In other words, sometimes security/capability gains more from perception than reality.
That was my explanation for awhile, but at some point the logic of the security arguments started to seem patently flawed. Also, judges seem to be ok with all sorts of judgments against domestic (state and federal) entities. It's almost as if when national security is mentioned, everything is thrown out the window. It's like a magic codeword or something and I don't understand the legal basis for it.
You make a good point re: investigatory capacity, but it seems even that is fuzzy and could maybe be handled through other means (aggregating time periods or using some kind of required delay).
I can’t imagine what “terrorist attack” would depend on knowing the exact number of requests, and fails if only a rounded number is known. Can you suggest any possible scenario?
Okay, you rule against the CIA and try sleeping well at night. Also, the judges that get these cases were undoubtedly picked (at some point of elevation in their career) for how "reasonable" they would be on these issues. You gotta remember that "national security" became a thing decades ago (Truman? Eisenhower?), so the system has gone through many several generations of selection pressure.
I guess you can level the blame at Truman. Eisenhower himself tried to warn us against going down this road, but the game had already been rigged against us.
I can accept many reasons for not revealing at the time, jeopardising a criminal investigation and that could cover national security investigations. So I can see how the judge has angled this.
What I can not understand or appreciate is that such requests can not be made public after said investigations are closed. Alas that approach is now also curtailed and would have to be raised as a separate avenue now. Which with this judgement, will now make that a little more harder to pursue. Been nice if the Judge had allowed that avenue of release, but that was not catered for in the scope of how this case was put forward.
Maybe if Twitter had approached with - what is a fair and reasonable amount of time until we can make such requests public? Or along those lines, then that avenue of release would of been within the judgment remit. Though I'm no lawyer, and may well that this approach they have taken is the best as it allows two bites of the cherry and an angle to appeal.
One just hopes the matter is not as clear cut and settled as it stands.
Twitter requested permission to reveal exact number of requests.
Companies are allowed to reveal gagged legal process in ranges of 1000, starting at 0, for six-month periods. Combined FISC and NSL requests can be revealed in ranges of 250.
Six years for summary judgment?!? Even if Twitter had won, it and the American public still would have lost through this endless delay. Something needs to be done about the glacial pace of federal litigation. Justice delayed is justice denied.
I wish there is a Twitter employee like Edward Snowden or Bradley Manning to leak those surveillance requests. Leaking documents seems to be the only way forward in the U.S.
I'd be very surprised if the courts would follow the extremely narrow definition of "speech" required for this to work. Is there any legal precedent otherwise?
You can’t force someone _to_ say something. You can only require they not say specific things. That’s the process by which the canary works. You set up a default state by which you say “this data has not been compromised”.
Nobody can legally force you to say that in the event it has been compromised. At least not yet, and if we get there, I really do fear for us.
Readit News
“National security” is a sham now and has been a sham every other time it has ever been used as a blanket override for human rights. Human rights to free expression are more important than any nation.
This is just government spies trying to cover their ass whilst they continue breaking the law. They've conveniently mixed it in with "but terrorism!" to ensure that they get to continue breaking the law.
The hypothetical situation above would be an invasion of someone’s privacy, and any smart terrorist group should be using a message service that is end-to-end encrypted (I’m not sure if Twitter DMs are), but let’s acknowledge that there is at least a plausible national security reason for the US government to request things from Twitter. These requests can be abused, but any social network that has a private messaging feature could be used to coordinate national security threats.
Edit for clarification: If the US government made one request to stop a terrorist plot that they had some specific knowledge of, that would be okay in my eyes. If the US government requested access to every Twitter user’s DMs to scan for possible threats, that’s not okay. It seems like Twitter is trying to disclose the number of requests so we can see for ourselves.
How can we tell that they just didn't straight up lie about whether the one request was to stop a terrorist plot, and not to gather intel on political foes or ex girlfriends? With no oversight and no one allowed to advocate for the rights of the individuals under surveillance, I am still not okay with one request to which they claim is stop a terrorist plot that they claim to have specific knowledge of, because there's no checks and balances to prevent them from simply lying.
"National security" is one of those thought-ending phrases that's thrown about to make people stop thinking critically about the matter. Terrorism is extremely bad, but it's not an existential risk, by its very definition.
> let’s acknowledge that there is at least a plausible national security reason for the US government to request things from Twitter
There is already a system that permits the government, investigating crimes, to compel information disclosure from a provider. It's called a search warrant, and requires a (very cursory, heh) review by a judge.
The government is asking not only for the power to skip that step (they've already taken that - that ship has sailed), but to prevent people from disclosing just how many times that's happened.
"National security" is not an umbrella term that means "bad things happening to USians", although I'm sure that the people using it in this instance as a cover story for widespread, illegal, extrajudicial espionage activity by the government would prefer that you think of it that way.
Remember: every legitimate NSL could have been a warrant. There's already a legal framework for how to investigate and prevent crimes, including mass murder. Additionally, anyone who's ever read a US federal warrant for anything involving sequences of bytes or information technology of any kind knows that the burden is already hilariously bad/low, and that they're very close to being rubber-stamp already. I personally have never seen a federal warrant (they're written under penalty of perjury) that didn't have naked factual errors in it, and I've only seen the ones that got approved/signed by a judge. There's a 100%, easily falsifiable, perjury rate in them in my experience. AFAIK none of the federal agents who have signed these sworn statements have ever faced any negative repercussions whatsoever for the falsehoods contained therein.
So, that hilariously low/token bar to a search warrant is what they're skipping in these cases. NSLs and related FISA warrants are extremely suspect. It makes sense that they would want to hide and obfuscate their use and existence. They're basically carte blanche for the government to inspect any communications they want - no justification required. The only reason they'd use them in place of a search warrant is because they are illegitimate.
This is absolutely not about terrorism, or threats. There's already a functioning-for-them (albeit broken in the legal checks-and-balances sense) system for that in the form of search and seizure warrants; they're trying to confuse you with the terrorism card.
This is a cover-up for an illegal power grab, nothing more.
Now I realize the client-side password could be saved and retrieved in the browser for ease of use, but I don't think that's going on.
Companies are allowed to reveal gagged legal process in ranges of 1000, starting at 0, for six-month periods. Combined FISC and NSL requests can be revealed in ranges of 250.
This just shows that Twitter is fighting all the way, as I mentioned in my other comment.
Unfortunately a site gets really big like twitter you just have to bow to the powers that be, or they'll make life hard for you.
OTOH, there probably are at least a few investigations into real cases of abusive or criminal activity. This may be getting used to watch what’s happening in other countries. There are a lot of possibilities for gray area questions that are neither clearly legal nor clearly illegal, and Uncle Sam knows that exposure could cause things to lean toward clearly illegal.
Let’s not pretend that privacy of Twitter messages is a human right, though. It’s a for-profit business that was already monitoring your traffic, and the government is asking for what Twitter already has access to. While the government may have a very different agenda, using Twitter doesn’t come with any privacy guarantees in the first place. Twitter’s not going to come arrest you, but they were never under a legal obligation to help protect you from the government or anyone else, and Twitter DMs are not private in a legal sense.
Alternatively they justify some actions with the war on drugs or fighting against pedophiles with: "think of the children".
If you promote strong encryption you might help drug dealers, pedophiles and terrorists.
There's also the preaching: "you shouldn't be worried of us spying if you have nothing to hide".
"We are empowered to stop terrorism." (or drugs, CP, etc)
"X can be used for terrorism."
"Therefore we are empowered to stop X."
Thus that agency gains power over all possible X. Like cash, any monetary transaction, photography, encryption, any messaging system whatsoever, or any number of other things that should make up the bedrock of a free society.
Are we? The article just mentioned grave or imminent harm to national security, not an existential threat.
Don't get it twisted: it's just a cover for domestic criminals inside the US government who have become accustomed to getting to break the law as much as they want for the last two decades. By making it seem special or important (it's not) they get support in suspending basic human rights such as publishing.
They have a point about a security risk but what they also sneak under that guise is their desire to hide intelligence capabilities and potential threats. It is not a risk if you don't have a specific threat or plan of attack,certainly not a risk to the nation. Vulnerabilty is not risk.
Revealing a request may thwart their investigation but the burden of proof that a specific threat that cannot be stopped if the information is public is required of them. Short of that, you can chit chat on the subway with a third cousin of a terrorist or something and they can say "if we can't monitor his DMs then we can't know if they are using code words to plan a terrorist attack"
On a different note, I think it is a failure if the US constitution to not clarify details like this. A constitutional convention to review and update the bill of rights and a few other things is long over due.
Well put.
(of course the answer is no, it is not)
Sliding into Twitter DMs and baiting fools to say they are going to use fake explosives aren't particularly productive things for the government to be doing, but there's probably a reasonable legal foundation for most of it.
Well, they can’t have it both ways. Twitter takes credit for the Arab Spring which was indeed an existential threat to several countries, at least their regimes, but still millions of ordinary citizens were impacted.
I don't trust US corporations at all and I thought that there was some kind of corporate conspiracy to discredit the US government to allow corporations to gain more power...
Now I'm thinking that the situation in the US is even worse than that; not only are corporations trying to discredit the government, but the government is working hard to discredit itself as well.
There are lots of recorded occasions of single governments doing this.
Governments are seen as having authorities that corporations do not, and thus are to be viewed much more skeptically as a general policy. Companies can't raid your house with machine guns, or toss you in jail for publishing an integer on your website.
It's not that corporations are trusted more; it's that the maximum possible damage by corporate evil is much, much, much, much less than the maximum possible damage by government evil. IBM participated in the holocaust, but they didn't directly cause it. That needed governmental authority. It's prudent to maintain several orders of magnitude higher suspicion and/or skepticism of government activities; corporations are basically sandboxed (e.g. no armies, for the most part) in this model.
Government and corporations are merging to become the same thing. I just assume any data I give to a corporation or government is public since they share your data anyway.
Of course it is.
Powerful and motivated interests have long been working to overturn basically every social program and every protection for labor, discrimination, environment, voting, etc. The goal is to "starve the beast," privatize everything to work toward a government that only runs the military and police, and to make sure it stays that way once it gets there.
Insidiously, they can win support by pushing the government to be more wasteful, more corrupt, less effective, more intrusive, more bumbling, and more frustrating to work with. The moment of victory is when they convince you that government fundamentally is a bad idea that cannot work well, all those taxes you are paying are just going to be squandered, and we'd be better off getting rid of the whole thing.
They have been at work for decades through groups like the Heritage Foundation, CATO institute, etc., to push this into the mainstream. They reached new heights of legitimacy with FOX news & co., and have made great progress with the Trump presidency so far.
Microsoft,Yahoo, Google, AOL, Skype and Apple all did more than legally required.
Dead Comment
It's great that twitter took a stand, but if that company failed everyone would easily replace it. Twitter is toy.
Well I guess it depends on what your definition of "toy" is. I think all popular social media platforms deserve careful consideration. They have a colossal impact.
It's more interesting to consider what, exactly, government surveillance can be used for.
And really, "James Bond" style villainous plots aren't what we _really_ need to worry about. The real weapon in these platforms is disinformation campaigns-- and the actual actors behind such campaigns (as well as anyone else who is paying money to influence any user) should be revealed to the public.
Like in this case, I fail to understand how releasing the number of requests would have any security bearing.
You make a good point re: investigatory capacity, but it seems even that is fuzzy and could maybe be handled through other means (aggregating time periods or using some kind of required delay).
[1] https://en.wikipedia.org/wiki/Pascal%27s_wager
And it's hard to challenge such law as unconstitutional because no right to privacy is explicitly written into the constitution.
Dead Comment
Dead Comment
What I can not understand or appreciate is that such requests can not be made public after said investigations are closed. Alas that approach is now also curtailed and would have to be raised as a separate avenue now. Which with this judgement, will now make that a little more harder to pursue. Been nice if the Judge had allowed that avenue of release, but that was not catered for in the scope of how this case was put forward.
Maybe if Twitter had approached with - what is a fair and reasonable amount of time until we can make such requests public? Or along those lines, then that avenue of release would of been within the judgment remit. Though I'm no lawyer, and may well that this approach they have taken is the best as it allows two bites of the cherry and an angle to appeal.
One just hopes the matter is not as clear cut and settled as it stands.
Companies are allowed to reveal gagged legal process in ranges of 1000, starting at 0, for six-month periods. Combined FISC and NSL requests can be revealed in ranges of 250.
https://www.rsync.net/resources/notices/canary.txt
... is now 14 years old. This was the first warrant canary[1] and, I am sad to say, one of the few remaining ones.
[1] https://en.wikipedia.org/wiki/Warrant_canary#Usage
Deleted Comment
Nobody can legally force you to say that in the event it has been compromised. At least not yet, and if we get there, I really do fear for us.
[From memory from some ancient HN thread]