The Side Channel

1:26 AM

@MaartenBodewes It's already -9. I really hope no one deletes it!

I want to see it get lower and lower.

@FutureSecurity I had someone claim that they could influence me to dox me from afar. I asked them how and they said they could screw with my computer by using some sort of astral magic bullshit. I told them to bias /dev/urandom to give more 0s than 1s to prove it. It's safe to say it never stopped passing dieharder. :D

Future Security

1:47 AM

My favorite RNG mystical claim: "Computers generate random numbers. Someone at ____ school monitors these numbers to see if human emotion has the power to influence things. Then the September 11 attacks happened and people all over the world immediately felt sympathy for the loss of (American) life. The people monitoring computer RNG output discovered that at that moment, random number generators all over the planet were not the same quality (somehow) as other random numbers." @forest

forest

ahahahaha

Link to claim? I vaguely recall having read that before.

Future Security

Not bookmarked, but I might be able to find the people that are "studying" this.

I may have heard it on the History Channel

forest

"The History Channel: Where the truth is History!"

Future Security

lol

I think I found it. Not the paraphrased quote but the school. noosphere.princeton.edu/measurement.html

forest

lol princeton?

Oh god this is hilarious.

This research is so bad it doesn't even belong on ViXra.

> We are driven by that evidence to infer that something like a "consciousness field" exists, and that intentions or emotional states which structure the field are conveyed as information that is absorbed into the distribution of output values of labile physical systems.

Future Security

1:56 AM

So many people actually believe stuff like this. I think its a case of "people treat it as legitimate so it must be true!"

forest

It's amazing how many people actually believe it.

Makes me wonder if I got in the wrong line of work.

Future Security

Or half true, because let's meet in the middle guys. The scientists can't be 100% correct. No one can!

forest

lol

"Any given theory or hypothesis has a 50% chance of being correct, because it can only be one of two possibilities: right or wrong!" - Every crazy person ever.

Future Security

Well scientists could be right, but so could group X, group Y, group Z, group W. So we should only trust scientists 20%.

I'm considering writing a userspace RNG which still relies Linux's getrandom and other platform's random number source. I'm having a hard time convincing people to not use homebrew RNGs.

With just enough unnecessary stuff added into the mix to satisfy these people. A library that combines OS random numbers with placebo RNG output.

forest

@FutureSecurity Why not use OpenSSL's CSPRNG?

Future Security

2:10 AM

Because I want the source code to be a single file

forest

What do you need from the RNG? E.g. reseeding/backtrack protection?

@FutureSecurity The very simplest would be ChaCha20 seeded from getrandom().

Future Security

I don't need it. But I want it to have anything someone needs.

forest

ah

I did something similar with the Linux kernel. They have two random APIs, one for secure random, another for less secure random integers (prandom_u32() etc). For some reason, the latter is used in places where it shouldn't be, so I modified the code to make it use get_random_bytes().

Future Security

I could XOR the placebo RNG output with bytes taken directing from getrandom and SecureRandom.

forest

But since I don't fully trust myself to do it correctly, I combined it with the regular LFSR-based PRNG, just in case, effectively XORing the crappy placebo LFSR with CRNG.

Future Security

2:14 AM

That would be useful. Too bad /dev/random superstition is still a problem.

forest

Indeed. I even saw something like that in the Tor Project where they were worried about adding random padding to cells because it would screw up urandom...

Now granted, their belief was that it would reduce the entropy estimate, not the actual entropy, but that behavior hasn't been true since early 4.x kernels.

Future Security

Ughhhhhhhhhhhhhhg

How old would that make those kernels?

forest

More than a year I think.

I have no sense of time.

Tor Project has a little problem with crypto theory.

Like they use HMAC-SHA3 instead of KMAC.

And they are considering replacing null pads in cells with random pads to make cryptanalysis against CTR harder, despite the fact that their CSPRNG uses CTR.

Future Security

Or someone has proposed it because it adds a covert channel or side channel attack.

forest

Nah, it's generated entirely on the client.

It was just a silly proposal from a dev who isn't huge into crypto.

The rest of the code and design is actually surprisingly high-quality.

Future Security

2:24 AM

That's good. What's SHA-3 used for? Isn't it still slow in software

forest

They're using it for the next-gen onion services (the ones that are longer than the existing .onion websites, which were so old they used RSA 1024).

But they're using HMAC with it which is just silly (SHA3 is not an M-D hash).

Future Security

Ya. At least it doesn't weaken security. Maybe using HMAC satisfies some low-knowledge high-paranoia users. Or funders.

forest

They just made a mistake in the design process.

And now it's baked into the specifications lol

Future Security

I found out that people really distrust the standard RNGs. They will cite Debian/Oracle, Android/SecureRandom, and some windows bug.

forest

I don't get why. Those are all crappy implementation issues.

Hell I could write a secure CSPRNG in my sleep.

Future Security

2:33 AM

What's worse is that they would rather use RNGs baked into the executable instead of dynamically linking to one.

forest

Indeed. Dynamic linking is a huge security benefit.

Problem in a library? Replace one file!

No need to recompile everything that uses it.

Future Security

I wonder how much damage these people actually do. Their software probably isn't used by many people. Some applications bad randomness isn't catastrophic. Like ad hoc password generators.

On the other hand, IOT.

forest

A lot. All it takes is to convince some obscure device to use it...

Then you get things like RSA BSafe.

Future Security

Hmm. Know any high profile examples in open source projects? Is there someone besides the OpenSSL team that needs to be shouted at?

forest

For RNGs in particular?

Future Security

2:42 AM

Let's say in general

forest

For RNGs at least, mbedtls can fall back to using HAVEGE in embedded devices. No idea how secure it is the way they use it, and I've seen a few embedded devices making TLS connections with randomness generated only from libhavege.

In general, oh god there is so much.

Like imlib2 is a clusterfuck of ugliness, a massive frontend to libjpeg/libpng/etc written so badly that every comment has a typo or two. Horribly insecure.

NSS (for TLS) is absolutely terrible in terms of being constant-time.

QEMU's sandbox is a joke. Check out their sandbox source code for lulz!

Firejail is one of the worst projects I've ever seen.

In general there is a lot of high-profile crappy security out there.

No one ever fucking looks at the Linux slab allocator for security.

There are probably like 5 people in the world who truly understand Linux mm.

Nginx is super popular but has horrible security because they care so much about performance that they cut out asserts and sanity checks! I could go on and on...

And Nginx considers lines of code so important that they practically obfuscate things by putting as much dense code in each line as possible.

@FutureSecurity Also routers. Every SOHO router until forever lol

Future Security

Oh. I like to stick my head in the sand and pretend routers and printers aren't a global problem.

forest

I like to hack routers and printers for the lulz

@FutureSecurity You know one of the things that scares me the most?

How insecure the core backbone of the internet is.

When you have friends who have literally compromised entire (small) countries nets...

When you realize that terabits of traffic are going through core routers using default passwords exposed to the internet, or using obfuscated "impossible to crack" security through obscurity hashes (which get RE'd in a day)...

Future Security

I suspect that various countries are sitting on really bad hacks but aren't using them based on the principle of mutually assured destruction.

forest

And shit like this:

This article makes the (pretty convincing) case that the multinational move against Huawei was coordinated among five eyes partners. https://amp.smh.com.au/business/companies/how-the-five-eyes-cooked-up-the-campaign-to-kill-huawei-20181213-p50m24.html

Tweeted by MalwareJake on December 15, 2018 at 5:50 PM

@FutureSecurity Yes, they are all sitting on "bad hacks". They don't use it simply because they don't want to reveal the exact bugs. But I can guarantee you that many countries have compromised entire ISPs in other countries.

It's the reason I see the IC as worthless subhuman scum.

Future Security

2:58 AM

Certainly. This isn't an arms race problem. It's theoretically possible that we could keep ourselves secure by using bug-free software. I don't like potential war being the only deterrent.

forest

@FutureSecurity IC doesn't just push for backdoors and insecure standards. They pay millions for high-quality exploits. That's why you have companies like Raytheon SI, Leidos, Vencore, VRL, Lockheed Martin, BAE, In-Q-Tel, Talos, Exodus, Zerodium...

And lol "Hacking Team" of course

@FutureSecurity You can definitely use and configure software such that it is secure even against a majority of nation-state exploits. The problem is that the infrastructure your whole internet depends on does not use such security.

Future Security

I'm not really familiar with that part of the internet. I do know that browsers, servers, and operating systems are perpetually buggy. I'm sure it's horrifying because the same business incentives and power incentives exist in all types of technology.

forest

yep

Browsers are horrific. OS kernels are nasty, though grsecurity is amazing.

Sadly they are customer-only at this point.

Chromium is actually decent, but still so complex than 0day exploit chains for it absolutely exist (RCE + sandbox escape + LPE or RCE + LPE in sandbox).

Future Security

I bet. And the rate in which dubious code is added on top of The Nightmare greatly outpaces our efforts to even analyze technology. Definitely more than our current ability to fix them.

forest

yup

Information security is in its infancy.

Future Security

3:12 AM

Luckily I guess. Even our modern societies are really young.

forest

I think fuzzing and static analysis are going to go a long way.

We just need people working on it. I mean Brad Spengler has already solved Spectre v1 and v2 for the kernel using a GCC plugin he created, but he doesn't release it. Anyone else could write it as well, but no one cares to.

Same with defeating integer overflows (another thing he did) and even ROP.

But people care more about security theater (cough KASLR cough) than real security.

Future Security

(We only learned that doctors should wash their hands between touching a cadaver and dealing with a patient some time in the last 100-200 years.) Things can get way way better if we strive for better. And avoid extinction.

And, yeah, it's pretty difficult not to understate how dangerous and unethical the IC is.

forest

Medical science is another thing that's very young.

I mean we practically only started doing real medicine during WW2.

And even now it's incredibly immature.

We basically just take an injury and squish things together in hopes that the body can naturally heal itself. Very primitive, yet it's the best we can do.

Future Security

The problem I see is that technology will make people so manipulable and make exploitation and suppression so easy that we're going to have to decide between utopian and dystopian. Without taking tech away from the powerful we can't stay in the middle for long.

forest

Exactly.

It's one reason I'm fine with being a blackhat. I pwn the devices that are already insecure. I get money, and maybe, just maybe, it'll convince people that the devices they are trusting are really not that secure and the companies lied to them.

Future Security

3:23 AM

And there are people ideologically opposed to the utopia no matter how much status quo or worse costs. Many Americans will are willing to pay much much more for their own health insurance to deny health insurance to the employees who touch their food.

forest

Yep

Future Security

Political compass. What do you think of it?

forest

Capitalists are against a utopia because they believe that the ability to make money is all that matters, even at the expense of others. They compare it to evolution by means of natural selection. I agree. Just like evolution, there are parasites and horrible, nasty tricks creatures have evolved to use to get ahead. Might makes right.

@FutureSecurity I think it's very simplified but not particularly bad.

I'm left libertarian by that compass (which is nice, since I identify as that).

I believe the government should be an entity that provides necessary services (education, negative income tax, emergency services, etc) and protects us from abuses. I do not think a government should deal with victimless crime.

In fact, I think the criminal justice system in general (especially in the USA) is horrific. You know, slavery is actually explicitly legal in the USA, as long as the slaves are prisoners? That's why American companies are moving from outsouring manufacturing from China to doing it in the USA, because there are slaves there.

@FutureSecurity Did you know that American state government employees are required by law to purchase their furniture (in their office) from the fruits of slave labor? From a company called "Federal Prison Industries" that runs for-profit prisons.

Future Security

I do know about that part of America's justice system, but haven't heard of the name of the company.

forest

Well that's just one of the companies, the one that state governments use.

But many corporations profit off of slave labor as well, e.g. Whole Foods, Honda, Walmart, Starbucks, etc. And a lot more, of course.

Whenever someone talks about e.g. slavery in Ancient Rome and talk about how evil it was to employ slavery on such a mass scale, and how fucking enlightened us Americans are that we do not employ slaves, I want to punch them in the fucking face.

Because they are WEARING CLOTHES made by slave labor.

Talking on their high horse about how enlightened their fucking barbaric country is.

Or people talking shit about the South in the American Civil War and how evil and racist they were... Let me tell you something, there are still black slaves being forced to work. Not enough slaves? Make certain drugs used by the black community illegal and watch as all your previously free citizens can be legally forced to make whatever you want and killed on the spot if they ever try to escape.

Future Security

3:32 AM

I agree. It is slave labor. Hillary Clinton's chosen successor wouldn't implement an order by the government to free certain prisoners because it would cost prisons more to employ people.

forest

mhm

The fact that slavery is still so common and institutionalized is repulsive.

@FutureSecurity It's a problem deeply ingrained in both government and corporations.

Future Security

It's hard to deny that it's a police state. In so many ways. Gov tied up with business and both tied to every other large institution.

forest

Exactly. And the same country bullies the rest of the world into behaving like them.

And yet in schools we teach that slavery is wrong and doesn't happen anymore...

When it's being done in an industrial scale.

"Waah it's so racist that a white person has dreadlocks!" Well met me tell you something, the war on drugs is disproportionately affecting the black community, leading to their imprisonment and industrialized slavery. That is racism.

So yeah if you couldn't tell, I disdain governments and modern corporations...

Future Security

Having health problems, I'm not very able to do something about these things. I'm aware of things like that.

forest

Being aware is enough to change your behavior even in the slightest ways to help fight against it. Simply not being ignorant is a big step forward.

I've always wondered if I could start a service for designing small embedded communication devices to be delivered to prisoners, like cell phones but more capable and discrete. Just something to make their lives a little less horrible.

The only reason I've held back is because I worry that it would be used by some of the worse criminals to e.g. order that someone who turned them in be killed, but I tend to think those folk, who are actually dangerous, are going to be able to do that anyway.

Future Security

3:46 AM

I've heard arguments for abolishing prisons. I acknowledge that they're probably right... There are the obvious problems that prison abolitionists are smart enough to have already thought about. It prisons definitely shouldn't exist as they are...

forest

I am a prison abolitionist myself. I think they should be re-worked completely.

The only purpose should be either for mental health care, or to hold someone who is a danger to society until a time that they are no longer considered dangerous.

So the only reason someone would have an extended stay is if it was painfully obvious that letting them out would result in harm to society (e.g. the serial killer Pedro Lopez, who said that the minute he was released he would kill again).

Future Security

There's the problem of people in clean suits with much more figurative blood on their hands. I want there to be a way to disincentive their anti-social behavior.

forest

For punishment, I recall a study showing that 5 years is the maximum amount of time someone should be imprisoned that minimizes recidivism, even for the absolute most severe crimes. Any longer, and it hardens them and makes them more dangerous.

Like Norris and Bittaker who tortured and murdered young girls. They would never have committed those crimes were they not locked up for more minor offenses.

@FutureSecurity The problem with them is that they get to pay a small fine.

And that's all. I mean sure, I don't think they should be punished with excessive imprisonment any more than a cannabis user should be, but I do think they should be disallowed from making the kind of decisions they are capable of now.

@FutureSecurity Look at it this way... The "hacking" I've done would put me away for decades in many countries. Yet a CEO can order "hacking back" and, even if their lobbying for its legality had failed, they would get no more than a fine.

Future Security

Maybe it's just a matter of prevention vs cure.

forest

The fucking FBI with rule 41 is now able to hack anyone, even non-Americans.

The same FBI that intentionally hosted child porn for weeks on their own servers.

If anyone else did that, they would be put away for life.

And fucking Americans are so bloodthirsty that they get mad when they see a country that does prisons right, like Norway, where the only purpose of a prison is to keep dangerous people out of society. Norway does not make their prisons torture chambers, and because of that, Americans flip out and call Norway evil.

Yes, evil for not making their prisons more dangerous and horrible than they are.

Future Security

3:55 AM

Our system is designed in such a way that it actively perpetuates inequality. Everyone having a good life would be seen as undesirable. The reason double standards like that can exist is because we have double standards with class too.

forest

You are 100% correct.

And when people do push for equality, it's for all the wrong things.

Women in programming? Everyone and their dog wants to push for it. Women operating mining equipment or working in the sewers? Total silence.

The real discrimination is against people committing victimless crimes.

Doing drugs, putting things in their body that their government dislikes.

Or downloading copyrighted material, being branded as a "thief". Yes, a thief who steals something yet does not deprive the original owner of said object. /s

@FutureSecurity Classism is a major issue in many parts of the world.

Like people are so afraid to talk about problems like black on black violence out of fear of being labeled racist. Well you know what, blacks do commit a lot of crimes. Do you know why? Because we put them in that position, and we are refusing to pay for better schools so they can lift themselves out of it!

We are putting them in a situation where getting a good education is hard, then we refuse to talk about it and find a way to fix it. So much for equality...

Future Security

I remember being told platitudes in school. Things like "Maybe you'll be the one to end world hunger." When those kids grow up and find a solution, "Well sorry. I meant one of you would find a way to end hunger and increase profits."

forest

yep

It sickens me how the job of schools nowadays is to make people ready for a job.

Grooming children for wage slavery.

That's the real reason why there's such a push in the USA to teach programming in schools. Not because it's so awesome and children should learn critical thinking, but because it makes them useful code monkeys to suck dry like a lemon until they're overworked and miserable. That's why they teach Java and JS instead of C.

Future Security

4:12 AM

My problem with the political compass is that I don't think the authoritarian left is as real as the other quadrants. If someone comes up with an example it's either fictional, exceptional, caricature, or ordinary human cognitive dissonance. No one is 100% anti authoritarian.

forest

It's not that it isn't real, it's that there aren't many existing examples.

Most authoritarian states today are authoritarian right.

Even "liberals" in the USA tend to be moderate to far right.

Future Security

I hate the compass because it creates false equivalence. I hate the left-right continuum because what's considered left or right isn't based on much besides shared opinions. If anything there should be an authority anti-authoritarian continuum and nothing else.

forest

Well like I said, it is a simplification.

But the spectrum between authoritarian and libertarian is well-defined and orthogonal to the spectrum between right and left.

Future Security

I think if you want to create an orthogonal compass then you're going to see things mainly only in two quadrants.

forest

It's true, at least in modern times.

Future Security

4:22 AM

Why is the anti-big-government pro-surveillance, pro-segregation, want to increase the power of police and soldiers, and ally with religious fundamentalists? It's because they're all authoritarian.

forest

But just because few people fit into certain quadrants does not mean that they do not exist or that people have never fit in there. It just means that there is not an equal distribution.

@FutureSecurity Well not necessarily. Look at modern Russia for example.

Future Security

In what way is modern Russia left-authoritarian or right-libertarian?

forest

Oh I meant in terms of not being anti-big-government.

While still being pro-surveillance and the like.

And right-libertarian would include "classical" libertarians (actually the first libertarians were left libertarians, but the term was hijacked by capitalists during WW2).

Left-authoritarian would include China or DPRK.

Future Security

Maybe. But I think anti-big-government is more marketing than reality.

What characteristic makes modern China leftist?

forest

State-controlled or influenced economy, very heavy regulation for income.

For example, they crack down extremely hard on gambling.

Note that leftist does not necessarily mean communist (which would imply an aversion to private property and capital as a whole). A leftist country can still make money.

DPRK is also an extreme example of that, but only because their socioeconomic ideology, Juche, is based directly on the older ideas of Leninist Communism.

Combined with a heavy serving of totalitarianism and censorship.

Future Security

4:29 AM

You can have state-"communism" or anarchic communism. Socialism doesn't need to be government organized. You can have markets in socialist or communist rule. It's not as simple as command economy vs unregulated economy.

forest

Oh I agree, and I support many socialist viewpoints myself. One problem I have with people who criticize communism (which I believe will be the way to go in a post-scarcity society, if we ever reach that point) is that they forget that communism is not automatically government-controlled. Whenever a government has absolute power, whether through fascism or communism, they will abuse that power. But it does not mean that communism, as an ideology, requires government control.

It's just that everyone forgets that communism's core idea is the abolition of private property and capital, and that socialism's core idea is giving workers control over the means of production. People tend to think "omg communism = socialism = USSR!!!"

Where "everyone" refers to Americans or otherwise westernized cultures. :P

Future Security

Some actual communists will insist that communism is inherently stateless. Under communism you can have "stuff" but you're supposed to not have factories.

forest

Indeed. It just depends on what family of communism you adhere to.

And socialism allows for factories, but only when it is controlled by the workers.

As opposed to some corrupt CEO that makes all the big decisions.

The reason I think communism may be ideal in a post-scarcity society is because the level of automation should make production so cheap that money is no longer an issue. It would ensure that, although there would be less jobs, you would not need a job to be able to live a comfortable life.

But currently, corporations are using automation and getting rid of workers, but they are not passing the savings they are making to the rest of society. So now there are less jobs, but the price for the same products remains the same!

The result is the distancing between economic classes.

It also results in even the middle-upper class living in an unstable situation. Many people with 5 figure incomes are living paycheck by paycheck due to the expensive nature of all the things they want, which puts them in a risky situation if there's ever an economic collapse. It makes being frugal so much more important.

Anyway, I think that's enough ranting for me for now...

Future Security

5:39 AM

\/\/\/\/\/\/
One time I wanted to stream an episode of Myth Busters. I didn't though because of the surveillance state. They make a lot of things go boom. But I had already used the same IP to listen to a podcast that was definitely "three hops" away from an anarchist and to visit cryptography related web pages.

I knew that people with tech related jobs were flagged for extra scrutiny. (I guess to make it easier for nsa to hack companies.) It felt like it was a combination of three things which I wouldn't like mentioned in court or on the news, so I had second thoughts.

Let me repost a link to "The Global Consciousness Project" for those who like pseudoscience. (It's RNG related.)

forest

6:30 AM

@FutureSecurity That's why I use Tor for everything.

Better to be known as somewhat paranoid than to have all my searches known.

kelalaka

8:03 AM

quantamagazine.org/…

forest

What's that even saying? I haven't heard that this was even a way to factor numbers, much less heard it was a threat to RSA.

Maeher

> RSA encryption works in a similar fashion. In simplified form, it goes like this: A user starts with a message and performs arithmetic on it that involves multiplication by a very large number (hundreds of digits long). The only way to decode the message is to find the prime factors of the resulting product.

They clearly have no fucking clue what they're talking about.

2

forest

lols

kelalaka

8:23 AM

I couldn't find the article.

forest

So like, I'm trying to memorize a complete cryptosystem which can be done (with effort) on paper). I selected RC4 and RSA, but my real problem is finding big primes. Is there any easier way to find a large prime than thousands of Miller Rabin tests?

The only technique I know is to select a large odd integer, then test it and add two if it isn't prime, then repeat. That takes... a lot of time by hand.

The only easy tricks I know to tell if it's composite is to check if it's even, check if it ends with a 5, and check if every separate number, added together, is divisible by 9.

forest

8:44 AM

Divisibility rule

A divisibility rule is a shorthand way of determining whether a given integer is divisible by a fixed divisor without performing the division, usually by examining its digits. Although there are divisibility tests for numbers in any radix, or base, and they are all different, this article presents rules and examples only for decimal, or base 10, numbers. Martin Gardner explained and popularized these rules in his September 1962 "Mathematical Games" column in Scientific American. == Divisibility rules for numbers 1–30 == The rules given below transform a given number into a generally sma...

I guess this will help lol

kelalaka

8:58 AM

@forest Mersenne primes? 3, 7, 31, 127, 8191, 131071

M_n=2^n-1,

forest

Nah regular old primes for doing RSA on paper. :P

Maeher

Why do you want to use RSA?

forest

It's easy enough to do by hand. I'm teaching my daughter (who's 10, so only basic arithmetic) about crypto. She's memorized RC4 and RSA (as in she knows how they work and can do them with a calculator or on paper).

My end-goal is to be able to communicate securely over an untrusted medium using nothing but pen and paper, because I think that's pretty cool.

I just don't want it to take 20 hours of math to find a two primes. :D

Maeher

El Gamal in a safe-prime setting seems more reasonable, since you can reuse the expensive parameter generation.

forest

But generating primes is the slow part (and making sure it's safe prime is even slower).

Maeher

9:06 AM

You only have to do that once

for RSA you have to do it for every key generation.

forest

Looking into that now.

Maarten Bodewes

Aargh, there it is again. Please help me code a certificate without using any libraries on StackOverflow. That together with a question that programs HMAC using strings. Sheesh, I'm really going to stop looking at SO all too often.

forest

Generating a large k for each message does mean a few thousand coin flips...

@Maeher I think I'll stick with RSA, since it doesn't require a new k, and because finding two 512-bit primes for RSA is about as easy as finding a single 1024-bit prime for ElGamal (checking about 355 numbers vs checking about 710 numbers).

Maarten Bodewes

9:26 AM

Hey, an example code page that does it right. After all that negativity about web pages I have to post a good one. Look at that nice yellow warning message. I love that:
https://rosettacode.org/wiki/MD5/Implementation

forest

Rosettacode is nice in that respect. :P

Maeher

@forest You don't actually have to find the prime by hand though. You can do it once on a computer and write it down. But whatever, my suggestion would be to just not attempt secure crypto on paper.

forest

It's for fun+learning, not for actual security. The goal is to do it with 0 computers.

Maarten Bodewes

Noooooooo, I'm out of coffee!

forest

Which means that generating it once with a computer won't work (and memorizing that many digits is a lot harder than memorizing an algorithm).

Maarten Bodewes

9:40 AM

Phew, found a sachet. Just enough to get me to the supermarket to get new beans.

Eeeeeeeeeeeeeeeewwww!

You've got to choose another mod, I think I have a breakdown :)

Rory can do it.

Maarten Bodewes

9:54 AM

"I need the same function to be used in the server side same encrypt mythology " .... uh, what? Oh, "methodology".

forest

Encryption mythology huh. Would that be Anubis? :P

Maarten Bodewes

Possibly, I was thinking methodism :)

Oh, sorry, "live" religion should not be called mythology. My apologies :P

PS I've got a bottle of water in my fridge. I'm particular to dry white wines.

forest

Why not? They are.

Maarten Bodewes

Because people take offense to it. I had a colleague who I explained what I thought of religion and he had a complete collapse soon after. I'm not sure it was related, I surely hope not, but lets not go there again.

I'll enjoy stupidity from a distance.

To my defense, I didn't even know that he was that religious.

forest

10:14 AM

People can take offense to anything. I'm offended by religion in general, but I don't tell religious people not to capitalize the word "god" if they so choose. :P

Rory Alsop

@MaartenBodewes I have had coffee, so... Yes :-)

forest

But I'm certainly going to call it what it is: a mythology. And come to think of it, mythology isn't even a politically incorrect term. I hear religious people discuss their mythology all the time without it implying that they're part of a cult (even if they are).

Maarten Bodewes

@forest We're great at looking back and laughing at the "stupid societies". I surely hope that there is a society left in 100 years that can look back at us.

forest

I must be a bit of a futurist. I look at our current society and laugh at its stupidity.

Maarten Bodewes

Yeah, well, I couldn't have made the previous remark if I didn't as well :)

forest

10:18 AM

:p

Rory Alsop

@forest yup. For all its technological advances, it's particularly stupid in many areas.

forest

It's just as stupid as it was a thousand years ago. Nowadays we can just spread our own little flavor of stupidity throughout the world at nearly the speed of light.

Maarten Bodewes

In the beginning of the internet I was hoping that we could spread our knowledge and humor all over the globe. And we do. I was just forgetting what would happen when the rest of society joined in. That's my own stupidity I guess. And yeah, I refuse to write "internet" with a capital letter as well.

forest

It was fun until the mid-2000s. Maybe around 2008.

M'vy

"Massification of usages"

Never goes you'd expect it to go

Maarten Bodewes

10:29 AM

Hmm, I'm missing the manual of my Coffee Machine.

M'vy

man c

c - genericised soft drink generator (ie coffee, coke etc)

This manual page documents the GNU version of c. c is a genericised soft drinks engine designed to provide a stream of the desired beverage on the standard output. By default, c makes coffee, white, no sugar. This is a historical choice based on the authors preferences. This behaviour is controlled by the options, and a wide range of other drinks are possible. Please note that not all combinations make sense, in taste terms, but that parsing by c extends only to syntax. Experiment at your own risk.

To be back on the topic at hand, you just have to see what happened to Facebook.

forest

> 5 sysadmin mug

hah

M'vy

There was a before and after massification

@forest ahah

forest

Honestly that's a reason I think Tor is so important. It helps resist the negative effects of massification by providing a secure space that cannot be censored.

So less pressure from the masses to force conformity.

M'vy

@forest Yeah. I don't know about that. I don't usually bend to social pressure anyway

And I'm not even sure the problem is censorship.

The problem is the considerable amount of insignificant things that are spread (if we're talking FaceBook), they lead to the burial of what's important and was the first concept of the platform, i.e. keep meaningful relation with the people you're far away from.

Now what is it? It's a repository of cat pictures, people sharing when they last went to the bathroom and political rants.

Maarten Bodewes

10:42 AM

I just use facebook to share my kayaking experiences. Anybody may know about them, and they are the only really interesting thing (if that).

I frickin' hate what they do with all the other information that they steal from me though.

To be fair: the web was not designed right in the first place. We would have done fine without 3rd party content in a web page.

M'vy

True

forest

Oh yeah, I get what you mean.

That's why I avoid regular social media...

Maarten Bodewes

11:44 AM

Hey, I managed to get my point across in the end. Kudo's for somebody very introvert to finally listen: crypto.stackexchange.com/a/65948/1172

OK, accepting your own answer because you wrote down an ECIES description is not terribly nice, but that's for another lesson.

2

forest

11:56 AM

lmao

Maarten Bodewes

12:35 PM

It's winter time. I'll drink a real drink: put in cup: cacao powder (1,5 tea spoon), raw sugarcane sugar (2 tea spoons), sniff of ginger root powder, (red hot) chili pepper, tiny bit of salt, stir. Put milk on stove. Add cold coffee cream to mixture and stir while milk is getting hot (don't boil). Add steaming hot milk to cup and mix again really well. Now that's chocolate drink. Not the crap you get nowadays. Adding pure chocolate chips to the mix makes it even better.

It could be an entry in Andy Tanenbaums book: "How to prepare your input".

Oh, I top my spoons to max height. You need somewhat more sugar than cacao and lots of both.

Maarten Bodewes

1:36 PM

Oh, my sometimes I hate SE, see this question here:
https://stackoverflow.com/questions/66066/what-is-the-best-way-to-implement-constants-in-java
The answers are shit and missing important information, e.g. to make your class final and uninstantiable. But that's not possible because comments nor answers are possible because it is locked. Spreading misinformation forever and ever.

tigrou

4:09 PM

Hello. I found this in bitwarden documentation (https://help.bitwarden.com/article/what-encryption-is-used/). "PBKDF2 SHA-256 is used to derive the encryption key from your master password. This key is then salted and hashed."
I am confused, why would you salt and hash encryption key (which is result of PBKDF2 ) ?

b degnan

4:33 PM

@tigrou I think the order in the documentation is off: crypto.stackexchange.com/questions/3484/pbkdf2-and-salt

SEJPM

4:45 PM

@tigrou I suppose somebody just misswrote the doc, as PBKDF2 does both of these things already

@forest remember that any deterministic public key encryption is trivially insecure

so you wil never get around the randomness during the encryption process

SEJPM

5:05 PM

@kelalaka why do you think this should be migrated to InfoSec.SE? It's a question about a crypto protocol and thus on-topic.

0

Q: Password recovery on Block-chain ( Hyperledger sawtooth in this case)

I am using bip39 specifications for 24-word mnemonic, Whenever user registers with his/her password, A Scrypt key is generated from this password and used to encrypt this mnemonic, The admin of the database in this way, can't decrypt it. Let's say a user forgot his/her password, there is no way ...

Future Security

A definition of religion that seems to fit with the distinction people draw between "religions" and mythology : A religion is a mythology large numbers of people would kill for or use to justify a war.

kelalaka

The problem with religion, most of the religious people are not well educated. And like every cult (political, fan, etc.) people on the top are power hungers.

Future Security

It excludes a lot of cults though. And cults do a lot of damage.

tigrou

@bdegnan and @SEJPM : thanks for reply. This is what I expected but wasn't sure (i'm no an expert in crypto)

Future Security

Ya. But I wouldn't saw uneducated. They are under-education, mis-educated, and indoctrinated.

kelalaka

5:22 PM

@SEJPM password recovery triggered me.

Future Security

And it's not just the top that wants power. People like to be assured that their race or ethnicity is more deserving than other people's. Employeers will lean on their employees. Husbands like control over wives and children. And even cults involve power. "Once you progress to the higher ranks you'll gain access to the secret to ____".

Woah! I saw the word block-chain in that question title and didn't even notice the password recovery part. Maybe it's like banner blindness people develop in response to internet ads. Password recovery is not a good thing, but it's probably worse with block chains....

I remember reading Bruce Schneier's blog when I saw he changed his advice on passwords from "write it on a piece of paper and store it in your wallet until you memorize it" to "Pick a phrase, use the first letter of each word, then randomly capitalize letters, make replacements like 'to' with '2'"

kelalaka

The randomly part is problematic, hard to remember

Future Security

5:51 PM

Add less than one bit of entropy per character in exchange for making your password exponentially hard to remember. It's terrible advice to tell people they should use a sentence as a password. And it's terrible advice to tell them to randomly capitalize and 1337speak letters.

The two together mean that people will have a hard time remembering passwords, will frequently need resets, and will choose weaker and weaker passwords consequently.

kelalaka

Ubikey?

Voting, last 2 hours

Future Security

6:22 PM

My advice would be to use single sign on in the workplace (only the workplace), write passwords to personal accounts in a notebook if your living space is secure and low risk. If not your wallet is potentially a good location to keep passwords. Store backups in a safe if you can.

I would further recommend that people use password managers if they're going to do most of their sign-ins on one device, but I like the paper advice first. There are going to be people who think password managers are too complicated or will want to stay low-tech. I suspect more people would comply with those "rules".

One password for the work place and require people to use it daily. They should be able to remember after a while and solves the "sticky note" problem.

And any high risk account's passwords should be memorized, with no copies on paper and perhaps none in a password manager.

I think if we ask people to memorize for many, many more passwords than they need to use on a daily basis then all their passwords will be weaker and more likely to be leaked.

And let grandpa use paper because his house is less likely to be broken into than the likelihood that someone guesses his grandkids's names or his birthday from their own PC.

kelalaka

7:12 PM

When does a question fall into twitter?

@FutureSecurity what about just a password manager on cellphone

Future Security

7:43 PM

@kelalaka I think that in a work place with desktop computers it would be inconvenient. Non-SSO TOTP would be an improvement over needing to memorize many different work passwiords. (And perhaps easier and more secure than SSO.) Personally I wouldn't store certain passwords on a cell phone. I don't log in to websites on my phone, but I don't see the harm besides theft if I did. I would not save any passwords on a phone that I wasn't comfortable with using on my phone.

I imagine that people keep track of their phone nearly as much as they do their wallet. You might be more or less secure depending on what you're trying to defend against. I like that the cell company cannot control my wallet (literally) and that paper isn't vulnerable to remote code execution. It depends on which thing is more easily broken into.

kelalaka

One can use a cheap cell phone, not connected to any internet, just for a password manager and keeps it in the office?

SEJPM

If anybody is looking for news on the election, please go to this other room:

2

December 2018 Cryptography Moderator …

Open Discussion for crypto.stackexchange.com/election/3

2

Future Security

It might be unrealistic, but just having the chip used to communicate with cell channels might be vulnerable still.

I'm not sure what phones are like now. I think I've read about a code system used to control pre-smart-phone cell phones. One of which is described as "reverse dial with no ring" or something similar. But it could be used to use someone's cellphone as a listening device.

Probably @forest knows what I'm talking about, if it's real. Also that there was one computer on smartphones to do the smart stuff and simpler one used to control radio in/out. The latter are proprietary, insecure, and have access to the bus.

I'll search later and see if there is a non-tabloid source I can find. (Speaking of) Does anyone know what happened following this story

Meta Crypto SE

8:16 PM

1

Q: December 2018 Community Moderator Election Results

Cryptography's third moderator election has come to a close, the votes have been tallied and the two new moderators are: They'll be joining the existing crew shortly—please thank them for volunteering, and share your assistance and advice with them as they learn the ropes! Also, please join...

discussion featured election

kelalaka

Congratulation @MaartenBodewes @EllaRose

5

Ella Rose

Thanks! Looking forward to giving back ^.^

kelalaka

Let's mark it here :P

b degnan

8:38 PM

@FutureSecurity Engineering Samples of most cell drivers have a programmable MAN, it's like a MAC on Ethernet. You can snoop anything, and spoof anything pretty easily. You cannot use this to passively listen to an inactive phone, just a call in progress.

Future Security

9:19 PM

I found hayes command exploits over usb no example of remote access

kelalaka

9:46 PM

Dial x silently, listen the environment?

b degnan

10:18 PM

As a follow up to my question about AES and leaking bits, I was going to look into but I have "gasp" hardware, which means that I'll have to tie up my equipment. Does anyone know if there's a software AES implementation out there written in the form that it would take in hardware?

Maarten Bodewes

@kelalaka Thanks, I'm sure we'll keep each other on their toes :) Beware, I may do even less on StackOverflow now so you may be needed over there!

b degnan

@FutureSecurity That's an interesting one. A lot of firmware is actually still updated in factories via old modem commands. My hardware still supports XMODEM to update the EEPROMs because it's small enough to be implemented in silicon, so I don't doubt that those AT commands were not really locked down because it's most like a hardware IC thing and not a software thing.

kelalaka

10:35 PM

@MaartenBodewes I still need some time to learn the libraries so will be around there.

December 2018 Cryptography Moderator …

Transcript for