Jonathan S. Rouach

1/ Jolt now supports zero-knowledge 🧵 This makes Jolt suitable for privacy applications — no SNARK recursion, no "wrapping", no sacrifice of transparency. Marginal increase in proof size (~3 KB). Prover time is essentially unchanged. a16zcrypto.com/posts/article/…

@zooko I've started worrying about breakthroughs not against cryptography, but on the contrary "them" finding new schemes and not letting us humans in on the finds..

Huh. I think Justin Drake is right. I think the threat of AI-powered math breakthroughs against modern cryptography is greater than the threat of quantum computer breakthroughs against modern cryptography. ⤵️

ok by popular demand we pushed Early-Bird to end of this week (Feb5), turns you don't like weekend deadlines.. :)

🚨 Early-Bird to ZKProof 8 in Rome ends on January 31st ! If you plan to join, May 9-10 (before Eurocrypt), now is the time :) Registration is here: zkproof.org/events/zkproof…

@VitalikButerin @lex_node I love what utxo.live/oracle did in Bitcoin :) Probably gamable by dedicated parties, not real time, but fully self contained..

We need better decentralized stablecoins. IMO three problems: 1. Ideally figure out an index to track that's better than USD price 2. Oracle design that's decentralized and is not capturable with a large pool of money 3. Solve the problem that staking yield is competition Tracking USD is fine short term, but imo part of the vision of nation state resilience should be independence even from that price ticker. On a 20 year timeline, well, what if it hyperinflates, even moderately? If you don't have (2), then you have to ensure cost of capture > protocol token market cap, which in turn implies protocol value extraction > discount rate, which is quite bad for users. This is a big part of why I constantly rail against financialized governance btw: it inherently has no defense/offense asymmetry, and so high levels of extraction are the only way to be stable. And, of course, it's a big part of why I refuse to give up on DAOs entirely. If you don't have (3), then again you have a few percent APY suboptimal return rates, which is quite bad. The possible paths to solving (3) [treat this as enumeration of the solution space, not endorsement] are basically: (i) reduce staking yield to like 0.2%, basically hobbyist level (ii) create a new category of staking which has yield almost as high as regular staking, but which does not have the same slashing risk (iii) figure out how to make slashable staking compatible with usability as collateral (does it mean that slashing risk somehow passes on to stablecoin and CDP holders, so both of those need to stake and trust the same delegate?) If you're going to try to reason through this in detail, remember that the "slashing risk" to guard against is *both* self-contradiction, *and* being on the wrong side of an inactivity leak, ie. engaging in a 51% censorship attack. In general, we think too much about the former and not enough about the latter. Also remember that a stablecoin cannot be secured with a fixed amount of ETH collateral; in the event of large drops you need to be able to handle rebalancing (though of course you could choose to partially drop this goal in a clever way, eg. if ETH price moves too much you stop earning staking yield until you take some other action)

it's increasingly obvious that Ethereum is a contrarian bet against most of what crypto VCs are betting on they're betting on: *gambling *CeDeFi *custodial stablecoins *'neo-banks' Ethereum is tripling down on disrupting power to enable sovereign individuals

"but here's the kicker" oh shush, your LLM is showing <Robots, robots everywhere> Humans where you at ffs

@genesobolev @simonw @charliermarsh but you need scientific discipline to resist throwing all your tests into the agentic loop.. by now we're so conditioned to "give the llm all the context you can", it's really hard to hold out the extra tests for CI validation.

@simonw @charliermarsh I'm doing the same! Treating coding agents as brute force optimizers and tests as constraints. Another good consideration is a different set of tests in CI, so they're outside the agentic dev loop to prevent overfitting to tests.

Wow, this has been really effective for working with Claude Code too

holy crap... this "mystery AI model" might crash the stock market it took some digging, but I figured out what this model is (it's not what you think)

1/ New post: Jolt now proves RISC-V programs with 64-bit registers (RV64IMAC), at speeds exceeding those we previously reported for 32-bit. 1.5M cycles/sec on a 32-core CPU, 500k cycles/sec on a MacBook. Here’s why this matters 🧵

Aaaaaaaaah @Zac_Aztec it works as a solana program! Generator is generic now and works for arbitrary circuits. Now we will land some Noir-Lang proofs on mainnet Mark my words, I will become the unofficial Chief ZK tooling Officer on solana

I shared a quick 5 min on what’s at the heart of Zcash starting at about 7hrs 2min in. @zooko immediately follows with some Zcash and @halfin lore. Grateful to @balajis for the opportunity. 🙏

We made a Bitcoin SPV contract on EVM chains economically possible by proving the whole Bitcoin history with Noir. SPV contract is available on Ethereum and Base. Proof of ~915,000 blocks took 109 hours, but can be verified by the smart contract instantly. We tried to aggregate all technical details here @olkurbatov/S1vEdGJile/edit" target="_blank" rel="nofollow noopener">hackmd.io/@olkurbatov/S1…

🚨New Bitcoin Whiteboard. We can do better than BitVM. 00:00 Bitcoin Bridges and ZK Verification 08:42 Addressing Critics of Rollups 13:28 Optimistic ZK Verification 16:56 BitVM1 vs 2 vs 3 25:45 Garbled Circuits 34:07 Glock: Future of Bitcoin Bridging

@vcorem @PeterSchiff @Vivek4real_ @pete_rizzo_ @BitcoinMagazine @AdamBLiv There's a model where you build such an energy facility, to power BTC mining and sell Gold and excess energy, maybe for useful pow for AI? :)

@PeterSchiff, I just want to make sure you didn't miss this. Manufactured "artificial" gold is coming sooner rather than later. Undistinguishable from "real" gold. Time to convert to Bitcoin, the only surviving hard money. @Vivek4real_ @pete_rizzo_ @BitcoinMagazine @AdamBLiv @adam3us

@vcorem @habitcoinist אז לפי החישוב שלך, איך זה ייגמר?

@habitcoinist משכנו את הדירה בפתח תקווה, קנו עם הכסף ביטקוין ברזל, משכנו את הביטקוין ברזל ( למשל ב AAVE ) וקנו MSTY. עם הדיבידנדים החודשיים, שלמו את המשכנתא על הדירה, החזירו את הריבית וקצת מהקרן על משכון הביטקוינים ב AAVE ועוד יישאר לכם עודף למחייה.

חלון הזמן בו אדם מהיישוב יכול לאסוף מטבע יחיד - אחד מ-21,000,000 שאי פעם יהיו -- נסגר מול עיננו לכולנו הייתה, ועדיין יש, הזדמנות לאסוף במחיר סביר, אבל העולם מתחיל להבין את יכולת הטכנולוגיה להגן עלינו מול החוב העולמי בשנים הקרובות, החל מ-2028 לדעתי, גם חצי מטבע יהיה נדיר 🫡

🆕 2025 in LLMs so far, illustrated by Pelicans on Bicycles youtube.com/watch?v=YpY83-… @simonw's keynote won our 2025 overall Best Speaker trophy, and was so well deserved! listen to the greatest highlights of 2025 in LLMs, Simon's explanation of the Lethal Trifecta in LLM security, and the release of PelicanBench!

@Yampeleg yes yes they added hooks so they want you to mess around with it already

is it ok to talk about reverse engineering claude code already or not yet? asking for a friend

I love privacypools.com so nice to be able to send ETH without exposing my account balance & tx history 😌 especially without having to KYC with some CEX that is then a honeypot just waiting to get hacked...

I just swapped a bit of shielded Zcash for Solana in @zashi_app via @NEARProtocol Seamless. Coming soon.

@jbaylina @ziskvm amazing - good luck!

Me, together with a core team of 7 developers, just spun off from Polygon to continue the ZisK project as an independent initiative. I’ll remain co-founder and advisor at Polygon, but my main focus now shifts to @ziskvm. Excited for this new era in the blockchain journey — we need this tech to bring truly decentralized values to mainstream adoption. 👇