Bruce21b ⓥ

Market of markets. That is why NEAR Intents is at Solana Breakpoint, building with Solana apps instead of against them.

After a week of the ICP patent drama, here is what I see. DFINITY started two lines the same week in 2021: a sealed patent application, and a public CC BY paper with the same method. For three years the public line was the only one anyone could see. NEAR built on it and cited it. Then DFINITY switched to the sealed line and asked for money. The product they built on the same patent holds about 240 BTC today. The dates, all public and linkable. Sep 30, 2021: DFINITY files a patent application on re-randomized presignatures. Applications are sealed until published, nobody outside the company can read them. Oct 5, 2021: five days later, the same two people, Jens Groth and Victor Shoup, DFINITY's cryptographers and the named inventors on the patent, put the same method on eprint under CC BY. Free for anyone to use, attribution only. Nov 13, 2023: a NEAR engineer opens a public ticket. He describes a known attack on ECDSA with derived keys and presignatures, and proposes the defense from Groth and Shoup's paper, citing it by name: "one technique proposed in [GroSho22] is to re-randomize presignatures". Merged in December. This is how cryptography is supposed to work: you do not invent your own, you take a published, peer-reviewed construction and say where it came from. The code running today still carries the comment "inspired by [GS21]" with a link to the paper, and the Trail of Bits audit describes it the same way. Nov 28, 2024: the application becomes public for the first time. Three years after filing, a year after NEAR cited the paper. Jul 28, 2026: patent granted. Sep 25, 2026: NEAR makes online presign the default, a path where the presignature is computed after the request and there is nothing stored to re-randomize. The PR description is about protocol safety, no patents mentioned. Oct 2026: DFINITY's founder calls it theft and asks for license fees back to the first day of use. Bottom line: two paths opened the same week. NEAR took the public one for three years and said so at every step. DFINITY waited until the sealed one was granted, then came for the readers of the public one.

Context for everyone else. Dom says NEAR took a signing method from DFINITY "without referencing them", knew about unpublished patent applications, and points at a 2024 PR as proof of a cover-up. Dom, you did not look. The source is named in three places. near/mpc_old issue #352, Nov 13, 2023: a NEAR engineer wrote up the attack, the fix, and the author: "one technique proposed in [GroSho22] is to re-randomize presignatures". The code that shipped has a comment "inspired by [GS21]" with a link to your paper. The Trail of Bits audit describes it the same way. Your application became public on Nov 28, 2024. The ticket is a year older. The paper has been on eprint under CC BY since 2021, published by your own people for anyone to read. So the facts are: public ticket, public credit, public code, a year before anyone outside DFINITY could see a claim. That is the opposite of hiding. You asked NEAR for "proper credits and references". They were there before you asked. Read the repo before accusing the people who wrote it.

Gemini: NEAR is live, deposits and withdrawals enabled. Gemini is the exchange that bet on Zcash before it was fashionable. Shielded ZEC withdrawals since 2020, the only major US exchange that offers them. Last month they rebuilt their Zcash node in Rust for the November upgrade that cuts block time from 75 to 25 seconds. The Winklevoss twins filed for a ZEC ETF. NEAR Intents is the main rail for ZEC outside its own chain: ZEC on Solana, ZEC in confidential swaps, and this week near(.)com's Universal Send, where you pay from USDC and the other side receives ZEC privately. So the step after listing NEAR: let users move ZEC in and out of Gemini from any chain, shielded on the Zcash side, through Intents. One integration instead of a node on every chain. Gemini has the fiat exit and the compliance, Intents has the routes. For clarity: nobody announced this, the pieces are just lying next to each other now.

I know some of you are scrolling this thread looking for my reply. Enjoy, this one is for you. > "Key enabling component is stolen DFINITY tech." The threshold ECDSA protocol in near/mpc is Cait-Sith by Lúcás Meier, an open-source MIT library, triples-based, not DFINITY's design. The repo is public, go read it. > "8 servers, nobody can get more info." The MPC network is 17 operators, threshold 11 of 17, names public: Bitwise, HashKey Cloud, Nansen, NEAR One, HOT, Aurora, Everstake, Luganodes, Stakin and more. It is in SVRN's SEC filing. His number is from a blog post dated May. > "The algorithm NEAR took." Additive key derivation is BIP32, Bitcoin, 2012. Presignatures predate DFINITY. The one thing from DFINITY is a fix from a paper Groth and Shoup put on eprint in 2021 and presented at Eurocrypt 2022. The paper's own abstract says these techniques were "used widely throughout the cryptocurrency space" before it. Citing a Eurocrypt paper is called research. Calling a published paper "proprietary, licensed for ICP only" is called marketing. There is no patent, I looked. > "Emergency downgrade from our algorithms." The PR he links was merged on September 25, a month before his thread. The issue behind it, from September 2, says the reason in plain words: three precomputation phases are hard to maintain, "a simpler system is better for trust and maintainability." It is a Cait-Sith variant with online presign. No DFINITY, no lawyers, no emergency. > "Two engineers mining the ICP repo." The GitHub link he gives is a NEP discussion about execution context in the runtime. Nothing to do with cryptography. > The fired employee story. No facts, just a grudge. I will leave it. So the summary: an open MIT library, a Bitcoin standard from 2012, and a public Eurocrypt paper. If that is theft, every wallet on earth owes DFINITY a check.

NEAR published the Chain Signatures update: 17 independent operators inside a TEE with proof checked on-chain. Great for NEAR. And it quietly makes OutLayer a lot stronger, so a few words on why. Trust in an AI agent has two floors. The lower floor is the key: who holds it, who can see it. The upper floor is behavior: what the agent did with the key, under whose rules, and can you check. NEAR just closed the lower floor. The key is split across 17 operators, each operator's piece lives inside encrypted hardware, and the chain verifies what code those machines run. Nobody holds the whole key. Not an operator, not NEAR, and you do not have to take anyone's word for it. OutLayer is the upper floor, and it sits directly on that key. The agent's wallet secret is derived by that same MPC network for an OutLayer enclave that proved its code on-chain. Nobody at OutLayer can read it. Then everything the agent does happens inside its own enclave, with its own proof: the policy you wrote, the limits, the confirmation step, the connector that places an order on Hyperliquid or Polymarket or pays an invoice through Mercury. The agent gets buttons, not keys. Put together: a key nobody holds, rules only the owner can change, and a proof for every run. Until this update one link in that chain was a promise, the MPC operators' servers. Now it is hardware and on-chain proof, all the way down. The one place that still trusts OutLayer is cross-chain swaps and withdrawals, which our coordinator assembles. In practice it means this: you can hand an agent real money and real accounts, and every action it takes comes with a proof that no one, not the agent, not OutLayer, not NEAR, could have faked. NEAR built the foundation. OutLayer built the floor you actually live on.

Hard to believe, but the hottest market on earth, AI compute, has no real exchange. Deals are OTC, by phone and PDF, like oil in 1983. It cannot go on an order book: GPUs are heterogeneous like oil, an H100 on InfiniBand in Virginia is not an H100 in Jakarta, and perishable like electricity, an idle hour at 3am is gone forever. Nothing with both properties has ever traded cleanly. So H200 hours cost what they did at launch and the whole market clears over Signal. This is a NEAR Intents shaped problem. You cannot post an order for a thing with no standard grade, but you can post a need: a thousand GPUs of roughly this shape, this region, these weeks. Solvers who have it, or can stitch it from five clouds, come back with a price. The biggest market without an exchange will get one, and it will look like near:native Intents, not Nasdaq.