e79

If you’re interested in computer science, have you ever looked at the Software Foundations course by UPenn? It follows a similar approach of having you build all sorts of fascinating math principles and constructions from the ground up. But then it keeps going, all the way up to formal methods of software analysis and verification.

https://softwarefoundations.cis.upenn.edu/

If you were to participate, could Google then refuse to hire you or generally do business with you?

I don’t think RNG or modular math were really the culprits here. PuTTY’s k value generation is deterministic and the biasing was caused by a mismatch of integer sizes and the resulting leading zeros. The offending operation is named mod, so that’s related to modular arithmetic, but the modulo (521 bits) was bigger than the SHA512 output (512 bits) from deterministic k generation. I linked earlier to a post where I break this down at the source code level.

Updated link with corrected title: https://ericrafaloff.com/biased-nonce-generation-in-putty-fo...

Good catch! I wrote the title before I had dug into the matter and forgot to update it. Thanks for pointing that out.

Any k generation and subsequent signature generation are going to be impacted.

I did a bit of a deep dive into this, in case anyone is interested. I think reading the code is a great way to understand _why_ this vulnerability happened:

https://ericrafaloff.com/your-putty-generated-nist-p-521-key...

I never got diagnosed with COVID, but I was almost certainly exposed. Mine developed suddenly during one of the early peaks in 2020. Unfortunately I was also producing music at the time with headphones, which I've considered the likely cause. But who knows -- maybe it was a combo of things. My stress was through the roof, I wasn't sleeping much, I put on a lot of weight, and I was exposing my ears to moderately loud sounds. Suddenly my brain just decided to keep looping a tone indefinitely.

At least treatments are starting to come out. It's nice to know that there is relief if I ever need it.

For anyone curious about how the exploit worked:

https://www.aon.com/cyber-solutions/aon_cyber_labs/an-analys...

I think my ideal setup is still the Push 2. You’ve got your mouse and keyboard for more tedious workflows and VST usage, and then the Push as an expressive platform to jam with and explore ideas on. Best of both worlds.

It seems like the Push standalone was designed for performers in mind? Would be pretty sweet to not have to bring a laptop up on stage with you, I’d imagine, especially if your setup is modular and otherwise DAW-less.

Agreed, but with the caveat that depressed people often struggle with making the changes they need the most. Similar thing with obesity. There often needs to be some kind of intervention, such as a medication, to offer enough relief for someone to break the cycle and start making changes. Willpower alone isn’t always realistic.

It’s fascinating how many potential modalities are at play with depression. Serotonin, inflammation, direction of brain signals. This suggests that depression may be a label that points to one of many underlying conditions, which could also explain why it’s so tricky to treat for some individuals.

This is not my experience in NYC at all. For example, waiting to see a sleep specialist took me about six months. An endocrinologist took about the same. In both cases I spent hours calling different practices and trying to find shorter wait times. My doctor says it’s still because of COVID and the shock it had to the medical system.

The contest approach wouldn’t necessarily hold rigor, because it doesn’t formally prove that all 2-MAXSAT problems can be solved using this algorithm. Just that one or more cherry-picked problems can. I think the paper really just needs to present actual proofs for the propositions it makes (as others have pointed out).

Too bad it’s Windows only. I would love an open source alternative to TouchDesigner.

This is a lot less depressing when you consider that vaccines have shown astounding results so far at preventing disease. COVID is a lot less scary if you’re extremely unlikely to get very sick from it. This article touches on that a bit, but it’s buried until the end.

There’s a lot of optimism to be had here, I think. Let’s not let perfect be the enemy of good — maybe we can’t eradicate COVID, but we can make it a lot more tolerable to share this planet with it.

I generally think of it as “cheap” inclusivity. Companies looking to cast the biggest virtual signal with the smallest amount of capital. On paper it looks nice enough, but dig a little deeper and you realize that it’s everything under the sun minus the most important part: training and hiring more women, people of color, and other minority groups in tech.

I’ve worked for companies that do this with mental health too. Everyone is burned out and unhappy? We’re bringing in a professional on workplace happiness! We’re partnering with non-profits! We’re doing everything! Except, you know, addressing the actual cause of the burn out and unhappiness. Because that requires a hard, sober look at our own behavior and wrongdoings. It’s so much easier to look outward instead.

I don’t identify as a minority in tech, so I won’t even pretend to understand exactly what that’s like. But it seems like being gaslit constantly. It sounds so painful and invalidating and exhausting and outright maddening to be told to look at all of this progress when you know they know it’s all the cheapest version of it and it’s mostly just for show.

The “smart guessing” approach reminds me a lot of SAT solving. Certain search problems become easier when you introduce randomness and clever heuristics to optimize and guide your guesses.

Edit: And here’s a link to the paper discussed in the article for anyone who is curious:

https://arxiv.org/abs/2007.10254

Re #3: I held off investing significant money into Bitcoin because I was convinced that the forces of the market would pull it back down to $10 where it belonged soon enough. I’ve been asking myself ever since why I felt so sure at the time.

As for GME, there’s fundamental analysis and financial experts telling us it is only a matter of time before the price free falls. But there’s a serious behavioral dissymmetry here. The market is rich with GME stock buyers who don’t care to listen to any of these traditional buy/sell signals. I’m not sure you’re wrong, but I’m personally a little bit less certain that the price will drop — at least soon — due to this reason alone. As this article points out, feedback loops can be extremely powerful. And I suspect they can play out over longer periods than we expect.

I hear this a lot: “it only kills old people” or “it only has a 1.5% death rate.” Many of these statements are followed by the suggestion that we should just let Covid run its course so we can be out of this situation sooner.

We still barely understand Covid. No public health official with an ounce of ethics and concern for their career is going to be in favor of letting a poorly understood virus just rip through entire communities. The death rate is only one part of the equation. As this article points out, scientists are now concerned about the growing number of cases in which severe symptoms stick around or even develop long after infection. When compared to the cold or flu.

We also just recently learned that antibodies only stick around for a few months.

There are reasons that public health officials are being so cautious. It’s not just to be “draconian” for the fun of it.

This has nothing to do with the original argument you referenced. Also, scientists are discovering more about Covid every day. What can’t yet be explained is a reflection of the rigorous scientific process being applied. Scientists are careful not to jump to conclusions. It can take years of research to confidently understand a complex virus like Covid.

Public health officials needs to adapt to the latest understanding and work with data and statistics accordingly. It’s extremely boring and un-sexy (i.e., it’s not like scientists in the movies making a sudden breakthrough), which is I suppose why so many jump to conspiracies. We crave clear and spectacular reasons for the things we fear and have no control over.

If you read the article, what makes this different is explained. The severity and number of different post-covid symptoms is alarming scientists. It’s unusual and therefore worth reporting.

Some journalists are without a doubt hyping this up, but to blanket long covid as some sort of media hoax can do a lot of harm. People aught to know that their symptoms may linger and even worsen far after two weeks. Censoring this just because you don’t like it or it scares you is harmful.

Yep you’re right. My bad. Hmmm... I still think my point makes sense. The “smash and grab” style attack fits given how noisy it was. People were wondering why they didn’t do something far more insidious like covertly gather everybody’s DMs and such. That’s not really feasible when you know your attack is going to get noticed fairly quickly.

A lot of people are asking “why a bitcoin scam?”

From what we know right now, targeted accounts had their emails and 2FA reset via an admin tool. These attacks were noisy, so the window of opportunity for the attacker was small. The attack was launched after hours, likely to limit the chance that the compromised Twitter employee would be around. So market manipulation wasn’t really a great option.

This was basically a “smash and grab” style attack, which makes sense given the noisy nature of the access. I wouldn’t be surprised if Twitter’s admin tool purposely doesn’t allow employees to silently access accounts.

Where is the author rolling their own crypto? They’re simply importing Node’s standard crypto library and using AES, a CSPRNG, and PBKDF2. Whether the usage is secure is another question, but I wouldn’t say this counts as “rolling your own crypto.”

It is technically light on theory. DSP is just a lot of applied math and it’s hard to hand wave it away if you really want to write software.

A book on music production or an introduction to audio engineering would probably offer a higher level overview.

I agree that they all have their advantages and disadvantages. My point was agreeing with the author that a proof assistant returning 0 for n/0 isn’t nonsensical. It’s debatable whether it is the best approach, but it is sound. The reasoning is clear if you consider R to be an inductive type, which many proof assistants including Lean do. Returning anything but an element of an inductive type would be wrong.

What would “halt and catch fire” look like for a proof assistant? If I’m trying to prove a theorem that performs division over the reals, would it then have to mechanically prove that there cannot exist any inputs that would result in division by zero? Isn’t that then itself a theorem that I’d have to explicitly define using tactics?

When I’m studying cryptography and mathematics, I find myself thinking visually and scribbling down little diagrams too. The ADEPT method [0] would suggest that this is one way to better absorb knowledge and form understanding. It’s really cool to see more content creators exploring visual communication.

[0]: https://betterexplained.com/articles/adept-method/

I think this behavior makes perfect sense when you view it through the lens of type theory. Functions that return a real number must always return a real number. Some function types are defined by something like ‘R, error’, and then it’s up to the caller to check ‘error’. Others throw exceptions that bubble up the call stack and may never return ‘R’, which in a sense breaks the type contract and has “exception” theory preceding type theory.

SSSS isn’t always used with fixed-size symmetric keys, in which case length can leak something important. But in practice it often is, since share size increases with message size and that can get unwieldy. So it isn’t usually an issue.