Youth corruptor. Mathematician. Father. The Fool. Player of Games.
Joined June 2009
- Tweets6.5K
- Following806
- Followers2.4K
- Likes7.4K
Pinned Tweet
A "writing" page to archive my writing, including favorite tweets and eventually a thread index: yanxzhang.com/writing
An Attack on Content-based Chunking… (joint with Boris Alexeev and Colin Parcival):
HN: news.ycombinator.com/item?id…
Tarsnap webpage: daemonology.net/blog/2025-03…
Paper: daemonology.net/blog/chunkin…
.. which means there would be roughly 50 moles in a L, so we should get about 13 moles instead. Multiplying by Avogadro, I get 13 x 6 x 10^{23} ~ 8 x 10^{24} molecules per glass. Since the volume of ocean was 1.4 x 10^{21} L^3, the ocean is 5.6 x 10^{21} glasses.
4/
So shouldn't that make 13 x Avagadros [sic] constant more like 8 x 10^24? That would be off by 10 again...
Current state: I'll assume Schrodinger meant "1000" until someone corrects me. Still confused at how AI had the same answer as Schrodinger at first...
9/END
In "What is Life," Schrodinger claims that if you could mark the molecules in a glass of water; then spread them uniformly into the oceans and sample a new glass, you would find about 100 of your marked molecules.
This exercise, somehow, turned into frustration at Society...
1/
This gives 12 * 6500^2 (km)^2 * 4km * 2/3 ~ 1.2 x 10^{18} m^3.
(sanity check: oceanservice.noaa.gov/facts/… gives 1.4 x 10^{18} m^3, so we're doing fine so far!)
Now a glass is about 250 mL (1/4 of a L, makes sense). How many molecules are in it? Well, H2O has 1+1+16 = 18 g/L,
3/
For estimation, we can assume the Earth is a sphere. I'll assume known that its radius is r = 4000 mi ~ 6500 km (which is itself a good exercise). I know that the ocean is about 4000m deep and 2/3 of the 4\pi r^2 surface area is ocean, and of course, for us, \pi = 3, so...
2/
On Representative Democracy in Voting
"In that Empire, the Art of Voting Theory attained such Perfection that the number of representatives of a single Province occupied the entirety of a City, and the number of representatives of the Empire, the entirety of a Province...
1/3
... In time, these Unrepresentative Voting Systems were no longer satisfactory. The Voting Theory Guilds struck a voting system of the Empire who votes for a set of representatives whose size was that of the Empire, and every person was represented by exactly one person...
2/3
Good luck @kanjun!
Explainer thread for Origami, joint work with @aard_k_vark (I'll try to stay high-level).
In interactive proofs, a Prover P wants to convince a Verifier V that a computation was done correctly. (blockchain people say "ZK", but the "zero-knowledge" is often unnecessary)
1/
With @krzhang, introducing Origami -- folding for Halo2 lookups. One more fold toward practical, efficient, inexpensive ZK proofs.
hackmd.io/@aardvark/rkHqa3NZ…
Overall, (at least speaking for myself) I think our contribution is a small but natural iteration as the ZK community collectively improves IVC.
It's also not foolproof, which is why we're putting it out here. I expect there to be lots of nuances to work out still!
10/10 END
Addendum 1: (thanks to @srinathtv) I made a mistake on Nova; Nova does have a quadratic relation and cross-terms, so it would have been correct to attribute that part to Nova instead and mainly credit Sangria for bringing folding to PLONK and sketching higher dimension gates.