I'm not an expert in infosec, but I do know a few things about cryptography, and I feel fairly confident in saying that this is misinformation.

I don't ascribe or suspect any ill-will, but I will caution that getting us to distrust tools that can help us to stay secure is a kind of propaganda that we can easily fall for. Partly because infosec is, as far as my lay understanding goes, is pretty fucking bleak!

https://jorts.horse/@AnarchoNinaWrites/115211389612661809

AnarchoNinaWrites (@[email protected])

I mean, whatever you think about "quantum computing" the fact is they are absolutely doing things in computing, right now, that are going to render all of your present encryption irrelevant. They CAN crunch the numbers now; or they're so close you should expect it very soon. And you're gonna tell me "okay but we'll just encrypt it more" and Imma say "what about all the shit you already typed under the old "secure" way that's still illegal?" Right. Stop thinking "forever" exists. It doesn't.

jorts.horse

Under all that bleakness, though, there's a few things that are just true, insofar as we understand the math. One of those is that public key cryptography *works*, unless we are wrong about some pretty fundamental assumptions about what kinds of math are easy and hard.

Math has given us an incredible tool that we can use to keep ourselves safe from adversaries with much, much, much more computational power.

We shouldn't give up that tool lightly.

Almost every attack we see on that cryptography (and here, I'll ask infosec people to please correct me) tries to work around that basic bit of math: making cryptography illegal, side-channel attacks, compromising your devices, spying on the people you're communicating with, or otherwise just making sure that you never get the chance to actually rely on good, well-designed cryptography.
If you don't trust cryptography, possibly because you don't trust the devices it's running on or the people you're communicating with, that's your wont. But please don't spread doomerism about it. Crytpography is an amazing tool, and as I said, we shouldn't give it up lightly.
@xgranade we agree with you, but we do also believe that the stuff you're saying needs to be held in dialectic with the point that safety is never perfect and never forever and pursuing it in absolute terms tends to cause other problems

@ireneista Zero question. The last thing I want to do is contribute to a false sense of security, or even just to the idea that full security is possible.

But I also don't want to contribute to a nihilistic belief that even things we know to be true as a matter of mathematical reasoning are suspect and thus not useful.

@ireneista It's from that perspective that I view the OP as well-intentioned but dangerous misinformation. Not perfectly, and never without asterisks, but I think it's important for those resisting fascism to internalize that cryptography *works* when done correctly.

Anyway, I suspect I'm preaching to the proverbial choir here, sorry...

@xgranade no, not at all, we wanted to talk it through for the benefit of anyone reading along, you know? we agree for sure that nihilism on this topic is incredibly dangerous

@ireneista Makes perfect sense, yeah! ♥

It's easy, and lord knows I do it in my own ways, to let the overwhelming sense of dread turn into hollow nihilism — but the least I can do is call it out when I see it, reassure that even if there's not a lot of hope, we do still have *some* tools we can use to resist, survive, and flourish. I strongly see cryptography as being one of those, even given the extreme difficulties of applying it in practice.

@xgranade for sure. every oppressive power structure has cracks made by its own contradictions. we must be the plants, growing in the cracks, prying the whole thing apart.
@xgranade because the useful perspective is one that kind of only emerges from taking those two simpler perspectives and really zoom in on how they interact, conversation is the best way we know of to explain it

@ireneista Absolutely. In the current case, I know *just enough* infosec to know everything is fucked, and I'm confident enough in my cryptography knowledge to know that it *doesn't have to be*. That resistance is possible, even given the overwhelming capabilities that are brought to bare against us.

Perhaps overly simplistic, but it is approaching midnight here, heh.

@xgranade and we're low on sleep as well, so yeah you're getting the abbreviated version of our sentiments too

but we strongly agree :)

@xgranade this reads to me like a rebranding of classic tech doomerism. it reads to me almost identical to the people who say "eh, they're all spying on everything we do anyway so there's no reason to bother with privacy-respecting software", a sentiment i never let stand when i encounter it.

honestly my best answer to this is "if this is true, why is the government spending so much effort on fighting encryption? why is chat control a thing? why do Signal keep getting pointless warrents?"

@Yuvalne @xgranade The other thing here is that as soon as an actor who could snoop on you covertly, performs an act to make you aware that the security is compromised everyone finds out and the exploit becomes useless.
So the only doomer scenarios to be concerned about really are those where the spys operate outside of the law, not those where e.g, the state wants to use legal apparatus against you - which tends to be the primary concern of the individuals in question.

@xgranade There are a lot of attacks on how that math gets used. Protocols can be extremely insecure even if they use the best primitives we know of. That said, we can also prove protocols secure, and people even do it these days.

I’d also consider some side-channel attacks part of cryptography.

But yes, I think your overall point holds.

@tryst How it's used, absolutely, and there are a rare few instances where as a society, we get the math wrong, but it looks very different than "everything is always broken, so why even try."

(As for side-channel attacks, agreed, which is why I listed those above explicitly.)

@xgranade there's a few things we remember from the crisis of confidence that happened around a decade ago which are not _quite_ these categories:

* there was some somewhat-wild speculation that nation states at the time could have had the necessary computational power to perform the pre-computation necessary to defeat diffie-hellman on the specific *prime field* well-known groups in common use at the time. TLS at the time also didn't tend to default to using forward secrecy. we've mitigated this by no longer really using prime field diffie hellman (also for speed), and by enabling forward secrecy. this wouldn't have involved novel mathematics at all, just more compute

* the biggest cultural change we saw since then was a *significantly* increased focus on implementation-error-resistant cryptography, which is the boundary between cryptography as mathematics and cryptography as something that engineers need to implement

@xgranade Yeah modern crypto is most likely bulletproof, the real concern are the devices used for decryption being compromised (all phones, chromebooks, and windows and apple computers are backdoored)

@xgranade
This is obvious disinfo, and IMO some of the most frustrating kind!
Setting aside the casual mixing of real threats and paranoid fantasy, the very structure of the argument is counter-productive garbage!
To throw away a tool because it might not work in the future, especially when there is no replacement, is a strict downgrade! It is worse than doing nothing. We need to keep fighting with the tools we have while we have them, not stop until some time in the future where doing nothing magically puts us in a better position with better tools that we will somehow get by not building them. This argument structure is just seeking comfort in giving up. If the NSA really does have some magic math that breaks all cryptography, make them actually use it!

Returning to the top layer of nonsense. Mixing true threats and harms with imagined exaggerations is the behavior I originally blocked this account for. It is pretty common on fedi though, Nina is by no means the only high profile account engaging this way. Half of the "journalists" here do it to get clicks. The whole point of this is to build a sense of authenticity and authority by claiming you are the only one who really knows how bad it is. This is extremely effective in left activist and similarly marginalized spaces because so many people are used to obvious harms being denied by corporate media and other liberal/fascist culture outlets. I think the people who exemplify it most clearly are the snake oil salesmen, see folks like Rob Braxman or Steve Gibson for some tech examples. Not everyone is doing it directly for financial gain though, plenty engage this way to farm engagement for their own credibility or to boost whatever real work they do, some even do it just for the ego.

I think a useful way to identify this kind of behavior is to watch how people pushing a narrative or "breaking a story" engage with technical or academic experts in whatever field. Look at Nina's replies and its clear she doesn't have a clue, even people exited to share the doomer perspective because they have expertise looking into the yawning void of computer security are correcting her on basic cryptography history.

@xgranade
Anyone who is engaging seriously with computer security as a front of political struggle will be talking with a lot of specificity about the vulnerabilities and failure modes of specific tools. People doing real organizing work understand that all tools are limited, and that you must work with what you have. The point of spreading knowledge about security flaws and potential hazards is to mitigate harm, either by fixing the issue or making people aware of how to avoid the circumstances where the hazard causes harm.

@xgranade

there’s two “theys” relevant to encryption:

the government: the US government has fired all its scientists and defunded all its research. “they” are not about to crack all cryptography.

Silicon Valley: they’re all busy holding bacchanalias until the LLM money runs out, at which point half of them will have forgotten how to program. “they” are not about to crack all cryptography.

@0xabad1dea @xgranade there are at least two governments other than the US with quality cryptographers however and one is even aligned with US foreign policy
@mcc
I guess China, who else?
@0xabad1dea @xgranade
@Blu @mcc @xgranade someone small, spectacularly well-funded for their size, and very busy with a genocidal disaster their prime minister refuses to back down from, so I’m not too worried about them having time for entertaining the Americans at scale
@0xabad1dea @Blu @xgranade I was wondering if @mcc was referring to the UK, which has a spectacularly well funded intelligence agency that specializes in signals intelligence, and domestic politics where all main parties and one insurgent party compete to serve the US ruling class at the expense of the domestic population. (I still don't think they can break encryption or will be able to any time soon.)

@0xabad1dea

Most of USA's governmental cryptology research is done under the aegis of NSA. It having "national security" in its name may have made it immune against the general purge of researchers.

@xgranade

@riley Trump has been open about hating them and expecting them to stab him in the back. They’re also notoriously queer friendly.
@0xabad1dea Oh. I guess there will be a bloom of new mathematics research in Europe soon, then.
@0xabad1dea @xgranade @whitequark nonbinary people are excluded from cryptography :(

@wxcafe @0xabad1dea @xgranade @whitequark

Non-binary people express their gender in qubits and so are ideally qualified to work on PQC.

@0xabad1dea @xgranade what about other governments?
@syn @xgranade also not about to crack all cryptography!
@0xabad1dea @xgranade I mean the US isn't the only nation on earth, China probably has some interest in cracking cryptography
@iagondiscord and they won’t. not in the paranoid omnipotence-ascribing “THEY…” way described.
@0xabad1dea that is true, I just think you'd be remiss to only talk about the US
@0xabad1dea @xgranade Oh yeah we should at this point be MUCH more scared that anyone who is clamimg they have broken encryption is lying and rhere's an LLM making shit up to back their accusations.

@0xabad1dea @xgranade

the government: the US government has fired all its scientists and defunded all its research. “they” are not about to crack all cryptography.

The Chinese government, in contrast, has been aggressively recruiting scientists for years and has silly amounts of spare money to spend on fundamental research.

When I was at Cambridge, I got Chinese universities trying to headhunt me with offers of guaranteed funding for quite large numbers of postdocs and PhD students. I would be shocked if they hadn't massively increased that recruitment programme and targeted US universities since January. If you have to work in a totalitarian state, picking the one that values science and research and wants to give you long-term funding seems like an easy choice.

@0xabad1dea @xgranade Cracking modern cryptography being kind of a weird thing when you actually put the math on what's required for it, I think people forget way too easily that it's measured in powers of two.

And it's quite visible with how few cryptographic algorithms have been cracked this last decade (like only that comes to mind is SHA1 and it was known weak in ~2007), what's been done instead these past decades is backdooring and cracking either the crypto implementation or something in it's address space.
@lanodan @0xabad1dea @xgranade MD5 and RC4 are also quite recent i think
@xarvos @0xabad1dea @xgranade MD5 got published in 1992, first collision done in 1996, and so broken in 2005 you could do collisions on a laptop.

Meanwhile best we got in SHA-1 is one collision in 2^60~2^63 operations, it's possible but extremely expensive for each collission.

@xgranade the general threat model. Supported by a bunch of orgs that trade in security, is that there's cryptography that's secure from Quantum and cryptography that will be broken as soon as Quantum happens. This seems to be true and has been just about to happen for as long as I've been aware of cryptography.

The theory of people is that governments are storing vast troves of data which they will decrypt as soon as quantum computers become viable.

This is a reasonable threat model, ish, but I'm not aware of any security program that hasn't been using "post-quantum encryption" for as long as its been available, so it's a mostly misinformed and outdated one.

@xgranade that account is invariably full of awful takes, so this is in keeping
@xgranade I think that kind of binary thinking is a bad practice when you have actual opsec concerns. Ignoring a widely used tool with a well-known security model because of some perceived wild risk like potential future quantum computers is likely to trap you into using something less audited that will likely end up having security flaws with much more near-term consequences.
@xgranade and on the quantum computing thing, there are very much already good tools to defend against such a potential future attack scenario today, and tools like signal already implement those. OP just seems poorly informed both about the offensive capabilities of near-term quantum computers (exactly zero) and the viability of current defenses (quite decent with literal decades of research backing them up).
@xgranade FWIW, I think I have OP blocked already for some other not very good takes in the past.
@jaseg I'm not terribly surprised. I don't follow OP, but saw that one boosted by someone absolutely legit, so I wanted to call out what I saw as well-intentioned but dangerous misinfo.
@xgranade This is mostly FUD, even when someone builds a QC big enough to usefully run Shor's and Grover's algorithms (decades at a minimum) they're not the crypto-destroying silver bullets that a lot of people seem to think they are