stephan.com

artist/scientist/technologist/inventor/dancer/dreamer/poet - not presently participating

[stephan.com @stephancom]

@AltCulture Every day, or at least every day I leave the house.

[Pope Leo XIV @Pontifex]

Dear friends, during some recent medical examinations, the doctors found a small, localized nodule in my lung. It was detected at a very early stage and is well-localized. The doctors believe that it is best to intervene, so, following their advice, I have decided to undergo surgery to remove it. The operation is already scheduled to take place in the coming days. I thank you for your spiritual closeness and entrust myself to your prayers.

[Cyan Banister @cyantist]

A week ago, I saw this graph and I couldn't sleep after. I realized that we've covered a huge swath of this investment wise already, so I decided to take this graph backwards and then forward.

Two words appeared in my mind:

MATTER COMPILERS

On Monday, on a random call, I was pitched a company doing just that and we invested. Yesterday they launched in the New York Times.

https://t.co/9x2dhomBFS

The world works in mysterious ways and stuff is about to get wildly fantastic.

[vitalik.eth @VitalikButerin]

I don't recommend anyone scramble to move their funds to new wallets today. But we should take the risks to cryptography from AI-accelerated math seriously, and minimize our exposure to not just quantum-vulnerable cryptography, but also potentially AI-vulnerable cryptography.

The core new area of risk from this viewpoint is, unfortunately, ML-DSA / FHE / lattices.

(and it's also another reason, along with quantum, why ECDSA might fall even faster than expected, hence the "fresh address" recommendation)

So far most people have been in the mode of thinking "elliptic curves broken, hashes safe, lattices safe". But there is a good chance that the concrete security of lattices will take serious hits from the next two years of AI math.

The basic threat model is: factoring is something that naively takes 2^(n/2) time, but over decades smart people have found and optimized number field sieves, and degraded that to 2^O(n^(1/3)), which is why RSA keys and signatures need to be ~400 bytes (and not 64 bytes). What if there are skeletons in the closet like that, both for elliptic curves and lattices, that we are simply not smart enough to discover - but bots soon will be?

This is a major part of the reason why for the past year ethereum's lean roadmap has been going in the "hash-only" direction: no lattices, no ML-DSA, no Falcon, no lattice-based commitments inside ZK proofs, etc. Signatures in lean ethereum are all hash-based, either WOTS or SPHINCS-.

For signatures and proofs, we already know how to go hash-only. The bigger challenge is for *public-key encryption* - and this goes far beyond blockchains. Secure communication, anonymizing protocols, lots of things need public-key encryption.

And unfortunately there are long-standing mathematical theorems showing why public-key encryption cannot be done with hashes alone. You have to have some kind of trapdoor object that has at least one form of usable "structure" - either group theory (incl. isogenies) or lattices or code-based or potentially in the future even more newfangled and spooky things (local mixing?). But for anything that has structure, you should assume that AI will make at least some progress in breaking that structure. Here, one reasonable inference is that if you want to make something plausibly long-term secure, multiply the key sizes by 10.

To me that's a very plausible world and something not at all extreme to predict. If AI will bring us 50 years of math in 2 years, then that 50 years of math may very plausibly include a "naive factoring -> GNFS" level of improvement to our ability to break lattices. In that world, lattices will still exist, but they will have to be significantly bigger to guarantee the same level of safety.

And at those new larger sizes, hash-based constructions will beat lattice-based constructions on concrete efficiency in every use case where hash-based constructions are possible at all.

Theoretically, of course it's possible that hashes are broken too (eg. P = NP would imply that). But I think P = NP is very unlikely. And intuitively, it's much more likely that a mathematical object has exactly no exploitable structure (like hashes are intended to), than that a mathematical object has exactly ~3 forms of exploitable structure (for elliptic curves: associativity, Schoof, pairings) and not some secret fourth form of structure we have not yet discovered that greatly degrades its security (for elliptic curves, ECDLP and pairing security). Similar for LWE, SVP, RLWE and the zoo of lattice problems.

For this reason, we do not yet see any reason to worry and start padding the byte size of hashes (if we start to worry more, we would pad the round count first before doing anything to the byte size).

Concrete TLDR, my own personal views:

* Hash-based > lattice-based, in those situations where hash-based is possible at all
* For anything lattice-based, be much more paranoid on param sizes. Remember that blockchains are only a small portion of the cryptography story; this point goes far beyond blockchains and applies to eg. access to websites, secure messaging, Tor / VPNs ...
* For privacy protocols, strongly favor NOT putting encrypted notes onchain. Instead, send them offchain through some third-party mechanism.
* If it's not difficult for you, keeping your funds in addresses which have not yet been used to make a transaction is a good idea. If it's easy for you, do it. **But be careful about migrations; I personally have lost more money in botched migrations than I have lost in all hacks combined**.
* For multisig wallets, doing confirmations offchain is better than onchain, because this way the signatures of signer wallets do not get exposed to the public, so if ECDSA falls to AI much faster than expected, at least the multisig "gracefully degrades" to a 1-of-1 where the 1 is whoever was gathering the signatures - a much better place to be than "anyone can take the money"
https://t.co/oVjwZog2lL

[stephan.com @stephancom]

@BNMT_1 @Mario_Sneh lol

[stephan.com @stephancom]

@AltCulture lol how much do I owe you for liking my comment? ;) hahaha

[stephan.com @stephancom]

@AltCulture Stop being so transactional

[stephan.com @stephancom]

@AltCulture I’d notice, and care.

[stephan.com @stephancom]

@EllasArchive1 If she didn’t say anything in the first call, how did the wife know it was a woman,

[stephan.com @stephancom]

@jan1183 @Immy_220 But it’s not impossible. No reason a woman couldn’t wear a larger watch. Perhaps it had once belonged to someone special.

[stephan.com @stephancom]

@grok do you agree with my method of extending the BRC map, and does the math look right?

[Marshall Mathers @Eminem]

I'm sorry we never got a chance to meet, but getting this note from Dolly really had a an impact on me. She definitely didn't have to take the time to do it. Even for a kid who mostly listened to rap music, she was an icon. A superstar in every sense. Rest easy Dolly. And thank you for the kind words.

[荒川和久／独身研究家／コラムニスト @wildriverpeace]

もうこれでいいじゃんと思ってる人多そう

[stephan.com @stephancom]

@Purepyo_ @farman4x Ditto

[stephan.com @stephancom]

@farman4x 12-8=4. 4/4=1. 4²=16.

Sixteen.

[stephan.com @stephancom]

Who wants to help Cards Against Humanity f*** with Elon Musk? https://t.co/6Y4eLFmfvW

[James Tate @JamesTate121]

Anthony Fauci's newly released diary is turning into a pretty rough own goal for the republicans who leaked it, because instead of exposing him as some deep state villain, it mostly shows a scientist working the phones, trusting the evidence, and quietly rolling his eyes at donald trump's chaos while trying to keep the country safe.
Senator Rand Paul released more than a thousand pages of Fauci's private notes from 2019 to 2022, and Fauci testifies before Paul's committee this week. Paul has framed the release around the idea that what Fauci wrote privately and what he told the country were two different stories.
Paul's biggest headline claim does not hold up well under scrutiny. He points to a diary entry where Fauci allegedly acknowledged early on that the virus did not originate at the Wuhan wet market, while publicly backing a paper that pointed to natural spillover from that market, and Paul is using the gap to argue Fauci covered for a lab leak.
But the entry Paul is citing actually shows Fauci complaining that conspiracy style letters from members of Congress were an amazing and unbelievable waste of time, which reads like frustration, not confession. A Dartmouth virologist told NPR that scientists being less certain in private than in public is normal, and that solid evidence exists for two separate viral lineages emerging, which actually supports natural spillover.
Paul has also repeatedly alleged, without evidence, that NIAID funded gain of function research in Wuhan caused the pandemic, a claim Fauci denies both publicly and in the supposedly damning private diary itself. If Fauci were really hiding something, you would expect the private version to admit it. Instead it says the same thing as the public one.
What actually comes through in the diary is a guy trying to steer a president who kept wandering off topic. Fauci repeatedly described trump's public comments as rambling, noting at one point that trump would drift into talking about TV ratings while Fauci tried to keep him on track. That is what a competent official sounds like when the man in charge cares more about ratings than case counts.
The timing of this outrage cycle is not an accident either. While cable news spends the week over a diary that mostly makes Fauci look diligent, RFK jr is overseeing the worst measles comeback in a generation. The CDC reported 2,318 measles cases in 2026 alone, more than all of 2025 and far beyond the 283 cases in 2024, with the number still climbing.
Not to mention the ever-increasing outbreak of diarrhea lettuce.
So, dig through Fauci's diary if you want. What you will mostly find is a professional doing his job under an erratic boss, and a senator whose greatest hits don't survive contact with their own footnotes. ~~VIA Exposing Political Corruption In The United States

[Millie Marconi @MillieMarconnni]

A Japanese engineer invented the QR code for one job, tracking car parts on a Toyota line, then his company chose to give the patent away for free, which is the only reason it ended up on every restaurant table on Earth.

His name is Masahiro Hara. The company was Denso Wave, a parts supplier owned by Toyota.

In 1992 the problem landed on his desk, and it was not glamorous. Workers on the factory floor were drowning.

Every car part had a barcode, but a barcode can only hold about twenty characters, so to track one component they had to stick five or ten barcodes on it.

A worker would stand there scanning a single part ten times in a row. Some of them were scanning close to a thousand barcodes a day. The job had stopped being about building cars and turned into pointing a scanner at stickers all day long.

And there was a second problem nobody upstairs cared about. This was a factory. Oil got on everything. A smudge of grease across a barcode and the whole thing became unreadable, and the line stopped.

Hara was asked to make the scanner faster. He looked at it for a while and realized the scanner was not the problem. The barcode itself was the ceiling. A line of black bars can only hold information going one direction, left to right.

He decided to build something that held information in two directions, up and down as well as across, so it could store hundreds of times more in the same little square.

Then came the part that sounds made up but is not.

Hara played Go on his lunch breaks, the old board game with black and white stones sitting on a grid. He was staring at the board one day and it clicked.

The grid of black and white stones was already a way to store information in two directions. That was the shape of his code.

But building the code was the easy half. The hard problem was speed, because the whole point was to be fast, and a scanner wastes most of its time just trying to figure out where the code is and which way it is turned.

The fix came to him on a train. He was looking out the window at buildings, and one building stood out from all the others because of its shape against the sky. That was the idea.

He put three little square targets in three corners of the code. The moment a scanner sees those three squares, it knows instantly where the code is and how it is rotated, even upside down, even at an angle.

Now here is the detail that shows how far he was willing to go. Those three corner squares only work if nothing else on the page looks like them.

If a magazine ad or a cardboard box happened to have the same black and white pattern nearby, the scanner would get confused and grab the wrong thing.

So Hara and his tiny two-person team went and surveyed printed material. Magazines. Flyers. Cardboard boxes. Piles of it, for days, reducing every picture down to its ratio of black to white area, hunting for the one ratio that almost never shows up in print anywhere. They found it. One to one to three to one to one.

That exact rhythm of black and white is baked into every corner square of every QR code on Earth, and it is there because it is the pattern the printed world almost never produces by accident.

Then he solved the oil.

He built the code so it carries a backup of its own information, spread mathematically across the whole square. You can tear off, smudge, or scratch out up to thirty percent of a QR code and it still scans perfectly, because the code rebuilds the missing piece from the copy it kept of itself.

A worker could get grease on a third of the label and the line would keep moving. This is the same math that lets a scratched CD still play and lets a spacecraft send data back across the solar system without asking to repeat itself.

He finished in 1994. He named it Quick Response, after what it does for the person using it, not after what it is.

And then Denso made the decision that actually mattered.

They held the patent. They could have charged a fee on every single scan, and given how many billions happen now, that would have made someone unimaginably rich.

Instead they announced they would not enforce their rights to collect royalties, and they published the specification openly so anyone could use it. Hara later said it was not even a big argument inside the company.

That one choice is the whole story. A code that costs nothing to use is a code everyone builds on. Airlines put it on tickets. Phone makers built readers into cameras.

Then a pandemic hit and the world needed a way to hand someone information without touching anything, and the free little square that a Toyota engineer built for greasy factory workers was suddenly on every menu, every payment, every door.

Hara still works there. He has said, more than once, that he never imagined it would spread this far, and that the part he is proudest of is that it got used to keep people safe.

The man built it to survive oil on a factory floor. It ended up surviving everything else too.

You have scanned his work a hundred times this year. Now you know whose it was.