Remco
@recmo
Followers
7K
Following
12K
Media
159
Statuses
4K
🪩 World | Math & Engineering @ 2π.com
Earth
Joined October 2009
I'm thrilled about this, we are extending libp2p with WebRTC support and compilation to WebAssembly. You will be able to run a full node in a browser tab! No need to install software!. FULL NODES IN THE BROWSER!. This is something I've wanted since before the bitcoin whitepaper.
Introducing 0x Mesh, a peer-to-peer communication network for the 0x ecosystem! 🌊💦.
16
71
345
Let's build free infinite flash loans straight into the token contracts and see what happens. Introducing FlashERC20!
20
42
292
I'm seeing too much discussion of theoretical ZKP performance and not enough about concrete performance. Lets compare 🍎≟🍏 and bench Polynomial Commitments!.
13
32
170
With EXP and LOG you can build a ton of useful functions, especially in finance (e.g. compound interest). Can be as cheap as 411 and 585 with some mathemagic!.
12
24
153
837,000,000,000,000 Dot Products in MPC. 52,882,698 World ID zkProofs Verified. 26,226 Merkle Tree zkRollup Proofs Verified. 93,116,233 User Transactions. 6,060,271 Verified Humans. 1 year. @worldcoin.
7
29
112
I wrote a TLDR on a number of signature schemes so I don't have to look them up again (or deal with exponent notation).
4
18
144
Skyscraper is 30x faster than Poseidon2 and comparable to the sha3 crate for Merkle trees. And this is a quick unoptimized implementation. Only assembly implementations optimized for short message batches are meaningfully faster. It's so good we even use it for PoW.
With Skyscraper, we show for the first time how hashing over large primes can achieve the same native speed as hashing over small primes in the ZK setting. This leads to an improvement of more than 50x compared to Poseidon 🚀 Check out the paper at #zk.
8
15
117
I'm starting up a team in 0x around ZKP-based layer two DEX scalability. Job description here: If you are interested in working with us on this cool tech or know anyone appropriate, please reach out!.
5
28
108
Groth16 is the most used and cheapest-to-verify-on-eth proof system. Here's my notes on the maths behind it:.
6
13
103
Never been more excited to read a codebase. I spend so much time thinking about what it would look like, finally get to see the real thing.
Stone Prover. Open Source.
4
7
97
Here's the secret to the optimization process
4
8
96
Plonky2 performance, recursion and Groth16 proof size/verification cost? You bet!
4
15
96
After a nomadic journey across chains, Worldcoin finds its forever home. Super exciting to be part of the Superchain working to scale without compromise. Even more exciting will be exploring what Proof-of-Humaness can do for fees and MEV to create a better experience for users.
6
3
91
It’s crazy that we don’t audit our food chain like we audit cryptographic software.
4
10
88
Why are you not working on the biggest problem? .Global governance is huge, yet you can start small and with a grant. There are exciting ideas in the @worldcoin Wave0 Grant and you can bring your own.With millions of users, you will build the largest onchain governance experiment
8
14
81
@worldcoin created 2 million WorldID accounts on L1. This is onle possible thanks to some efficient zkProof batch insertions we developed for Semaphore. But how can you trust these proofs? @bkiepuszewski explains how you can verify them yourself!.
So you have heard about zkProofs, how cool they are and how using them Ethereum can verify computational integrity of some off-chain programs (ranging from custom apps to full-blown Layer2 zkEVMs). But how do we know what is that program that is being verified ? 🧵👇.
20
66
72
If you want to know how modern compilers work, shelf the dragon book and read the cairo repo:. * Incremental computation.* Infallible parsers and concrete syntax trees.* Formatter, diagnostics and language server.* Multiple intermediate languages. Awesome @LiorGoldberg2 et al!.
The day has come: a first look into Cairo 1.0, fully open-source 🦀. This is a big step towards open-sourcing the StarkNet stack. You can now get familiar with the new syntax, compile and run simple programs locally. #StarkNet support is coming soon.
3
10
82
This is what US is like every time for non-citizens.
I’ve seen some long immigration lines before, but nothing like this. The video is so long this app won’t even let me post the whole thing. Welcome to Brussels indeed 🇧🇪🫡🛂
14
3
76
The Ethereum community was spending several hundreds of thousands of dollars PER DAY for L1 calldata that disappeared almost overnight with 4844. Imagine if this was instead directed at shipping 4844 faster. How do we tackle this going forward?.
17
5
70
This is the one that is going to bring mobile proving to the masses!. - Transparent setup: no giant downloads. - Small proofs: no large uploads. - Memory/CPU friendly: RS encoding + Merkle tree. - Recursion friendly: few field operations in the verifier.
WHIR 🌪️: a IOP of proximity and multilinear polynomial commitment scheme with exceptionally fast verification time. Joint work with @GalArnon42, Alessandro Chiesa and Eylon Yogev. 📚: 📄: 🧑🏻💻:
3
8
72
Speedrun the Worldcoin Tech Tree, Any%!. Calling all people working on zk-Identity and L2 scaling to apply.
4
14
62
We need a modern cross-platform open source lib for processing electronic travel document. Last three weeks I've been reading through a library of standards and implemented most of it here:. Tested in a dozen countries already, expect hundreds more soon!.
3
5
71
You can have Groth16 and universal setup if it’s a recursive verifier taking a transparent/universal system as public input. Groth16 is still hard to beat in calldata size, and that’s basically the only thing that matters onchain right now.
One of the many things that we need to acceeeeeeelerate is abolishing groth16. Per-application trusted setups are just not ok in the 2020s. Universal setup at the minimum, ideally no setup at all. This requires big improvements on infra and standardization for newer SNARK algos.
8
4
70
If I prompt my seed phrase to DALL·E 2, how hard would it be to recover my wallet from the image?. (Also, someone with API access please generate a seed phrase image).
4
2
68
Love the codebase! It's almost entirely Yul. Let's make Yul a first-class standalone smart contract language.
Introducing Seaport, a brand new web3 marketplace protocol for safely and efficiently buying and selling NFTs. With an emphasis on flexibility and optimizations, Seaport has been built to support new and evolving use-cases for where NFTs are heading.
4
1
65
Out of curiosity I took a geiger counter on a transatlantic flight. Got to ≈2.45μSv/h where Warsaw ground level is more like 0.08μSv/h. That's 30× background levels. Cosmic rays are real.
6
2
64
Plonk+FRI over a small field with useful properties is the way to go for flexibility and performance. I'm reading through the code and the implementation is also rock solid and highly optimized. Love it!.
We're excited to announce Plonky2, an implementation of PLONK+FRI which is focused on fast recursion. After experimenting with several approaches, we've reached a level of performance that we're happy with, with recursive proofs taking ~170 ms on a Macbook Pro.
0
7
65
Ethereum, the world's computer, has a ~200kHz processor and less than 1MB memory. But it can not be stopped!.
1
4
61
Noir gut! Who’s in Chiang Mai next week? I want to build a WHIR/SpartanNIZK backend for it for mobile proving.
pretty excited for @NoirLang getting popular with more devs as we wrap 2024 and enter 2025, @Zac_Aztec and team have been cooking on devex & perf. was just reading thru zkemail in Noir -- it really just reads like rust (repo below)!
12
3
62
PR for generic elliptic curve linear combinations, hoping to get this in the next hardfork. It solves a very large class of problems. I'm excited by the support from great minds at @StarkWareLtd @the_matter_labs @aztecprotocol et al.
1
18
59
𝔾₁ and 𝔾₂ point decompression can be done for 2.4k and 7.6k gas making it economical on L2s where calldata is relatively expensive. Should lead to good fee savings when using Worldcoin PoP.
6
3
54
It's not an EVM compatible chain if it doen't have an etherscan clone and a gas price oracle.
1
3
54
Never do this. I've chased down problems to incorrectly bot-closed issues so many times. I'm not going to re-open and add comments just to have it closed again in a month.
🤖 This handy action flags issues and PRs that go untouched for a certain amount of time. Try it out in your repo today and start closing out those stale issues!
1
3
53
Shamir Secret Sharing as used in MPC is one of the few things we _know_ to be unbreakable by quantum computers. Most 'quantum-resistant' cryptography just means we don't know how to break it (yet).
world id AMPC encryption is quantum secure.
4
4
54
We would have had AGI by now already if Python had a half-decent package manager.
1
3
53
After >2 years looking into all biometric template protection schemes and evaluating them on performance and security, the state of art was simply not good enough. Then @TACEO_IO showed us some recent advances in SMPC, and a great *new* solution quickly emerged. Very excited!.
Introducing SMPC.
3
4
53
Blockchains are great, but the real reason cryptography is moving so fast now is additive notation:. Then: gᶭᵆ = gᶜᶞ gᶡ gᶣᵜᶣᵜᶼᵔᶣᵙ.Now: R = s G + e B.
I wrote a TLDR on a number of signature schemes so I don't have to look them up again (or deal with exponent notation).
3
4
50
Circumference of the planets in light-seconds:. Mercury: 50ms.Venus: 125ms.Earth: 130ms.Mars: 70ms.Jupiter: 1.5s.Saturn: 1.2s.Uranus: 0.5s.Neptune: 0.5s. Global video calls are a small-planet phenomena.
9
2
52
In high school I made a sport of writing the shortest answer that would get full marks. I didn't realize it at the time, but this has been an incredibly useful life skill. Not only do you communicate more effectively, you also learn to extract the essence of an argument.
learning how to say something in 30 seconds that takes most people 5 minutes is a big unlock.
6
1
51
I'm thrilled to do the closing keynote at #EthCC on scaling Blockchains, DApps and @0xProject.
1
4
51
Excited about this!.The wallet has Semaphore's Circom circuit compiled to native mobile using Arkworks' Groth16 implementation and the WASM witness generator. Thank you @PrivacyScaling, @identhree & @arkworks_rs 🙏. Let's see much more ZK adoption wallet side!.
1/ Introducing World App. Built by @tfh_technology, its simple yet powerful tools are designed to be accessible to everyone. Use it to prove personhood w/ World ID, claim Worldcoin where available & send digital money anywhere for free.
2
4
51
I cleaned up and pushed the Yul parser I wrote some time ago. It uses the same design as rust-analyzer, so it is a lossless infallible parser. Great for a yul language server!. You can try it with `cargo install yulc`. Or read the libs source here:
Interest check: YUL→EVM compiler, but in Rust with pattern matching and all that goodness for rewrite rules. Should be easy to combine it with solc frontend to compile Solidity.
2
10
47
Optimism has been on the forefront of scaling since forever and with Superchain we are entering an exciting new phase! Looking forward to work alongside great teams to create great infra and build the billion people chain!.
World ID & World App are coming to the @OptimismFND Mainnet 🪩 🔴.
7
5
43
@_prestwich instead everyone should just run a SNARK inside FHE. > yes we can solve the problem! * ** ***. * depends on correctness of a 6 day old preprint.** prototype ready for audit in 74 months.** can only handle numbers less than 57 in the Grothendieck prime field.
2
4
47
Customizable Constraint Systems in a tweet. Operators ranging over sets leads to much cleaner notation.
3
0
44
zkEVM seemed far fetched not that long ago. Now there are many, all unique and incredible feats of engineering. Massive congrats to all the teams!.
0
5
45
Thinking about zk-verified-builds. @SuccinctLabs has anyone run solc in SP1? Can it be done?.
6
0
46
🏛️ Ruint v1.0.0 released!. All the traits and methods you know from 🦀 Rust uint primitives, but for 🐘 ANY number of bits!. Great 📚 Doc, 🧪 Test and ⏱️ Benchmark coverage. Ready for use, expect many more features added during v1 series.
3
4
44
Had a great evening chatting with @chriswaclawek and @RaoulGMI about humans, identity, privacy and blockchains at the @CIBlockchain meetup!
2
7
38
Today I wrote a tiny implementation of LLaMA in pure numpy (inspired by PicoGPT). It also does some hacks to memmap the weights, so you can run large models regardless of GPU/RAM (but very slowly).
1
5
43
The Modulus team is amazing to work with!.They're just as comfortable doing on high level protocol theory crafting as they are diving into low level implementation magic. Super thrilled to see them become an integral part of World as we venture full steam into zk-identity future!.
1
5
44
The future of EVM L2s will be based on Reth. Reth ecosystem is shipping fast, has several talented teams and is focused on solving the hard real world problems.
We have made great strides in the world of blockchain scaling in the last 10 years. It’s time to go further. Reth’s path to 1 gigagas per second, and beyond…where we present our roadmap for scaling Ethereum.
0
0
44
Imagine a world in which people can have as many legal identities as they please. What could go wrong?.With anonymous proof of person hood we can still have progressive taxation, welfare, and even a criminal justice system without ever linking identities.
7
1
42
Just published 📦v0.2.0 of the const-generic Uint lib. It is. 🏛️ now called ruint, and.🦀 works on stable with a.🛋️ good type ergonomic compromise, and an.🚀 even better uint! macro, and is also.✈️ compatible with a ton of crates!.
1
1
42
My post on scaling DEXs!.
Learn how we are working towards evolving the 0x tech stack to scale decentralized exchange for a tokenized world! 📚🌍.
3
2
39
@gakonst @stonecoldpat0 @technocrypto @sidiropo @ercwl @jtriley_eth I don’t think people realize how ridiculously large PoW got. It simply dwarfs all other compute on this planet. It makes training big AI models look like grains of sand. FOR TEN TPS. All eyes are on hacks and scammers, but really crypto should be lynched over this.
1
4
40
BN254 point compression got audited and merged into Gnark!.
Worldcoin project contributors have developed & open sourced an optimization technique that modifies the Golang contract template of the Groth16 proof verifier in the gnark library to optimize the verifier for both L1 & L2 gas consumption.
0
3
32
Summoning Danksharding, hope this works!. 🕯. 🕯 🕯. 🕯 🕯. 🕯 Danksharding 🕯. 🕯 🕯. 🕯 🕯. 🕯.
🕯️.
0
2
39
How ‘bout we rename “Coinbase, Inc” to “Fee Recipient, Inc”?.
I think it's time to rename the EVM opcode `COINBASE` to `FEERECIPIENT`.
4
1
40
Crypto twitter, meet Edgar. Edgar is like Etherscan but for CeFi, from our friends at the SEC.
3
4
36
New 0x network economics model mathematically derived from design goals by Peter Zeitz. I'm learning so much from him, the paper is a great read!.
0x Roadmap 2019 (part 4) — Proposal for Stake-based Liquidity Incentives 🌊.
0
5
39
@transmissions11 @onbjerg to be a pro is to be a noob forever. srlsly, the set of things you know you don't know will always grow faster than the set of things you know.
1
5
35
@gakonst @stonecoldpat0 @technocrypto @sidiropo @ercwl @jtriley_eth Energy spend on PoW is more than 2x consumption of all US datacenters combined. And it gets us like 10 tps. 10 tps in zkEVM takes maybe a couple dozen GPUs. These numbers are not even in the same universe. There’s like an Avogadro constant between them.
2
3
36
This is seriously impressive!.Having worked on this a bit way back, it's super hard to solve this in real time, with fresh data, taking into consideration slippage functions, fees, many tokens, many pools, anticipated MEV manipulation, and probably lots more.
1
6
37
While we are working on zk-proofs and decentralized state machines, tradfi is still struggling to move on from EBCDIC. The tech gap is incredible.
0
5
31
My attempt at improving on the fastest PCS (Winterfell's FRI). Note that you can do FFTs on Goldilocks for any root of unity efficiently. Doesn't need to be 2ⁿ. Also K12 hash beats Blake3 on M1.
2
2
27
I predict floating rock desk toys will become a profitable market nearly immediately, even with the tiny flecks we see so far. We can scale up from there.
If LK-99 turns out to be the real deal it’s manufacture is still unreliable. It’ll be decades before it becomes a viable product. Rare-earth barium copper oxide high temperature superconductors, discovered in the 80’s only became available as long tapes in last decade.
2
1
30
With capital abundant and talent scarce the smart VCs (VTs?) know how to maintain an edge. Congrats @transmissions11 and @paradigm!.
I'm thrilled to announce that I've joined @paradigm as a Research Engineer, to help @matthuang, @FEhrsam, and the rest of the amazing team build the world’s best crypto-native investment firm.
0
1
35
Pretty sure in absolute numbers @worldcoin already does the most compute in MPC and ZK (client-side, server-side & on-chain verification). And it's only getting started!.Next couple quarter should see MPC scaled up to GPU clusters and very large mobile proofs.
Despite all the baseless discussion to the contrary, I suspect @worldcoin is going to be one of the largest companies pushing privacy technologies (ZK, MPC) at massive scale.
2
6
35
Thank you! Modern cryptography is truly amazing.
@LouisAnslow IMO they actually have been doing quite a good job of taking the privacy critiques seriously and designing their system to be more and more data-minimal. Modern cryptography is truly amazing.
1
0
36
I contributed these earlier to Solmate, but v7 is taking some time. Meanwhile Paradigm is hoarding that unpublished sweet gas alpha in ArtGobblers and 0xMonaco! No fair!. Time to set these babies free!.
2
0
35
@BeeblbroxZephod @DJSnM @Serkis_92 No promises made to Russia were broken. The only promise broken are those made to Ukraine.
1
0
36
And soon we’ll measure it in “per second”. We are serious about scaling chains! Always looking for talented individuals and teams to join forces and build the billion people chain.
day 3 of @worldcoin launch, crazy lines around the world. one person getting verified every 8 seconds now.
5
2
33
PSA: 'Quantum secure' cryptography just means we don't know yet if and how to break it with a quantum computer. This is similar to 'cryptographicly secure' and classical computers, except we understand classical much better.
"This gives a polynomial time quantum algorithm for solving LWE.".Is this real?
1
4
35
PSA: Biometrics ≠ private keys.
@mike_connor Not how biometrics work. Anyone can lift your fingerprints from a drinking glass. It would be a terrible secret key. What authenticates is that you have that fingerprint on a *real life finger*. Without the lifeness check there is no authentication. Data alone not enough.
3
4
26
2
4
29
@huff_language The real challenge is, how do you exploit this without getting sniped by MEV bots.
4
0
33
This should give big improvements in Plonky2 style recursion.
Say hello to STIR 🥣!.STIR is an IOPP for RS which, compared to FRI, has shorter arguments (~2x) and a faster verifier (~1.2x) that performs fewer hashes (~2x). Joint work with Gal Arnon, Alessandro Chiesa and Eylon Yogev. -
0
1
30