Yulia

@buchmanster By making us focus on the minority of events that are evil, it plants and spreads the evil in our heads

feels like social media is some kind of accelerated evil-digestion machine forcing us to constantly digest greater and greater evils and continuously accept the world is more evil than we thought, while emphasizing that there's not much we can do about it but shout into the void

If you are in Berlin today, don't forget to drink your electrolytes. And check the ingredients - if it contains magnesium oxide as the only source of magnesium, find a new favorite brand. Magnesium oxide is cheap but poorly absorbed.

@badcryptobitch There is also no incentive to explain subprotocols in all detail if they introduce significant costs into the main protocol

If only academics weren't so imprecise when explaining subprotocols, there wouldn't be so many bugs in software 😅 If you are going to half-ass explaining an important subprotocol due to space, at least cite it so I can go read the complete protocol!

I just want to use my own app on ios is that too much to ask for (yes)

@colludingnode I'm socialost

Yes, I’m a socialist (love going outside and socializing with my friends)

Remember that cleartext transactions are not vulnerable to the deanonymization threat of quantum computers. That is why I prefer cleartext - it is safer

@yuvadm I noticed the change of the icon only after I opened the app and it screamed at me "hey, we are celebrating 20 years of spotify". Maybe the app didn't scream at you and didn't create all these "stories" or maybe you didn't open it, but can't say spotify kept it secret

@vveiln if I have no idea what the occasion is, we're just left with a horrible looking icon

Why do people hate it so much? It isn't the most "aesthetic", but it is appropriate for the occasion, no?

Honestly, to me MPC feels like nothing compared by complexity to ZK, so if MPC is so fragile, what should we expect of ZK? Interactivity makes MPC worse and it is hard to compare different categories, but then let's just think of co-SNARKs and other interactive augmentations

@apriori0x Coping mechanism is a strategy

ironic detachment is not a strategy it's a coping mechanism

@apriori0x I doubt you can serve the society without understanding what it needs (how being insane, desperate, excited, articulate, honest, or credible are enough for it?) or the ability to productively communicate with those who help you build and integrate

You don't have to be insane or desperate. You just need to be genuinely excited about the thing, articulate, honest and credible. If you're excited about the thing its not even work. You will just be motivated to figure it out and do it.

Mimic AI writing style so that nobody realises they are reading your actual thoughts

@apriori0x I can stop using claude anytime I just choose not to

if the people around you are impacted by claude and you are using daily you are not in full control. 99.9999% have weak cogsec. just embrace it -- emotionally healthier.

This imaginary scenario creates a lot of other issues and questions to answer. But perhaps it isn't the worst idea to run this experiment in our imagination while building our products. Would I trust the product I'm building?

Imagine if everyone developing products would be obliged to use them (for some time). It would certainly increase the degree of responsibility, but on the other hand, everyone involved would have to consent to releasing the product, which sounds like a nightmare

@cryptodavidw If you love writing, how can it be pointless? The point is to do something you love, AI can't possibly be better at it. People still do a lot of things that are automated for ages not because they are better at it than machines (although many are), but because it is fulfilling

I feel like AI has taken a lot of my drive to write. I used to LOVE writing so much, as one can attest from my blog. Since LLMs have become so good at writing it feels so pointless to write. It's not so special anymore. You don't stand out. Worse, AI writes better than me. What's the point anymore? I understand Lee Sedol.

@zksecurityXYZ It gives a perfect secrecy and provable security tropes vibe. Who fills the gap jumping from the spec proven in lean to what the agent actually implements? Do we trust the agent to interpret the spec? Who writes the spec?

Every programming language → compiles to assembly Every proof system → can encode semantics So: assembly + proofs = terminal object of software dev I think we reached it. Curious if people disagree ↓

This post ends software development. No Rust. No C. No compilers. Just AI writing RISC-V assembly + Lean proofs. If it compiles and proves, it’s correct. This is the final form. 👇

Kinda related point raised in the context of Qday Prize failure: confirming something works the way we imagine is less informative than confirming something doesn't work some other way algassert.com/post/2601

There is nothing special about cryptography when everything is well, it is just excessive math. The magic happens in adversarial settings, so it is more important to understand how a cryptographic system tolerates various attacks

@veorq How about this one: 🟥: elliptic-curve crypto until 2050 but vibe-coded 🟦: PQC but secure implementation and formally verified

🟥 red button: elliptic-curve crypto until 2050, secure implementation, side channel-safe, meaningfully formally verified etc. 🟦 blue button: PQC (your favorite) but newly vibe-coded, at least matches the test vectors no button for hybrid 😈

I factored the number RSA1024-1 using my home-built QPU stack; alarming sign that RSA1024 will soon be broken. I'm choosing Full Disclosure, in the interest of transparency and Science advancement: gist.github.com/veorq/25bee6ef… Non-ZK proof that the correct RSA1024 was used: #RSA-1024" target="_blank" rel="nofollow noopener">en.wikipedia.org/w/index.php?ti… @yuvadm your move