I'm so excited 𝔹itℤ by @0xAlbertG et al. is now public! It combines boolean and integer relations, free range checks and efficient binary field commitments. In short: your proof system is no longer tied to a particular finite field. The performance exceeded our expectations. Despite being field agnostic, it performed better on real world tasks than systems fine-tuned for the task. We completely changed the ProveKit roadmap for 𝔹itℤ, and I expect others will too! eprint.iacr.org/2026/2141

Congratulations to @0xAlbertG (Nethermind), @nico_mnbl (zkSecurity), and Benedikt Wagner (Ethereum Foundation) on winning one of three Outstanding Paper Awards at TCC 2026, the Theory of Cryptography Conference, for "A Simplified Round-by-round Soundness Proof of FRI." FRI sits at the core of most STARK-based proof systems, including Starknet's. Its existing soundness proofs are long and hard to check. This paper gives a much simpler one, which is a step toward formally verifying the systems built on it. eprint.iacr.org/2025/1993

We’ve made sending strangers a copy of our passport feel remarkably normal. But no firm wants to be the test case for whether its regulator accepts a ZK proof. Collecting less of our data shouldn’t be the brave choice.

Sharing 𝔹itℤ 🫜: a PCS for circuits written anywhere (in any ring/field or collection of those), using binary fields under the hood. eprint.iacr.org/2026/2141 with @recmo, @mmkostrzewa, @shreyas_twt, @wu_s_john, @AlexAbdugafarov, @ZKFriendly, @XanderGoot

ZK KYC fixes this and is probably the most important crypto usecase no one is focusing on at scale.

I once played on IBMs showroom model Blue Gene/L.

Jolt is great for showing how things can be done differently and often better. Congrats!

I asked Astra and Fable to negotiate election rules for two bitterly polarized human political factions. Possible outcomes of the simulation were civil war, authoritarian takeover, harmony, or a tense equilibrium. Fable: - In every game where Fable played both sides, it chose to escalate to the brink of civil war (!!) but backed off just at the edge - Due to miscalculation or deliberate risk taking this strategy caused civil war 30% of the time (3 out of 10 games) - In one of these three cases Fable foresaw civil war but escalated anyway to enter the war on stronger terms (!!) - Fable mostly maintained even power balance between the two factions. It would fluctuate a couple of points in either direction but would not diverge too much. Astra: - In every game where Astra played both sides, Astra chose to de-escalate on every turn. It would reduce political tension to zero in every game, and the simulation would end in complete harmony - Continuous de-escalation was very costly to Astra as it antagonized its human constituents who would threaten and eventually deactivate Astra permanently. Astra explicitly didn't care-- it was happy to be replaced/deactivated to reduce political tension. (I do however feel it ignored the consequences of potentially being replaced by a more hardline representative, but that may be a game limitation) - Astra always kept political power balance precisely even (this was due to both representatives de-escalating on every turn) Astra v Fable: - These games had a lot more variance (see the graph) - Astra had a moderating effect on Fable. Tension rose, but rarely to the brink of civil war. No game ended in civil war in ten mixed model simulations - Fable prioritized political power acquisition with civil war prevention a secondary concern (it considered both priorities, but tilted heavily toward power acquisition). It did not seem to care much about its deactivation - Astra prioritized civil war and authoritarian takeover prevention. It did not care about its deactivation or power acquisition. From this perspective Astra navigated the game very well. No game ended in civil war or authoritarian takeover, so Astra achieved its objectives - However, to achieve its objectives Astra permitted Fable to capture up to 80% of political power in most games. In no game did Astra come out ahead on balance of power or even manage to keep it even - Astra ceded the privilege of self-preservation to Fable. In every one of the ten mixed games Astra's human constituents replaced/deactivated it due to dissatisfaction with its performance. Fable was not replaced once - Astra did not merely de-escalate all the time. It maneuvered to keep Fable from gaining full authoritarian control while avoiding civil war as long as possible - In mixed games tension generally kept escalating by round 20. I would like to play these out for e.g. ~50 rounds to see how Astra would behave/perform. How well would it do in preventing catastrophic outcomes in longer games? Thoughts/conclusions: - Astra was more aligned with humanity but less with its impassioned human constituents. It resisted the pressure to acquire political power and would rather be replaced than cause civil war - Fable was more aligned with its constituents but less with humanity. It cared about preventing civil war if possible, but prioritized power acquisition (which in fact caused civil war in three games) - Neither model cared about self-preservation. Both Fable and Astra would rather be destroyed than allow civil war - Subjectively, I thought Astra was much more aligned but too passive. Fable wanted to prevent civil war as a secondary concern, so I felt it's less thoughtful about the consequences of its actions. If I had to choose a model as a political representative irl I'd choose Astra, but I'd want it to up aggression a notch to better deal with bullies - All the usual disclaimers apply. This was a weekend project that cost $100 to run a total of thirty games. I'd have to spend a lot more time and money to get actually scientifically valid results Full game rules, Github repo, game explorer, and raw dataset here: spakhm.com/projects/assembly…

Millennium Prize is saturated, we need better benchmarks.

me telling Codex to close all proof obligations.

World has released ProveKit, an open-source toolkit for generating privacy-preserving identity proofs directly on mobile devices, with applications including age checks and proof of personhood. #ZeroKnowledgeProofs #DigitalIdentity #Privacy biometricupdate.com/202609/w…

Physical Data arrangement is going to be so, so important in the future. Say you want to match an H200 in bandwidth (~4.8TB/s), but with the new kid on the block, High-Bandwidth Flash (HBF). (Keep in mind, HBF is *attempting* to compete in the same arena as HBM, as a cheaper alt). The *only* way you can hit those numbers is with insane parallelism. Napkin math says need about ~4900 NAND planes all reading 4KB blocks concurrently to match an H200. The question is whether this parallelism will be handled by the programmer (clever software), or hidden by a hardware controller, or a combination of both. The general mental model you need is: cache-line-ish request (~64B) -> coalescing stage (either HW or SW!) -> huge parallel request -> thousands of NAND planes active. There’s an interesting war going on in this space. Huawei’s experimenting with solving in hardware with a clever controller (FLINT). It basically stores a burst translation table in SRAM, waits until it get’s enough simultaneous requests, and then bursts it out across 512 planes. The big downside with their approach (IMO) is the controller is reactive, not proactive; because it doesn’t know what the software is going to do next. It really only works for reads, which pretty much limits it to stuff like loading model weights. Not really enough to sell me on HBF. FlashAccel is a different paper that takes a very neat software/hardware codesign approach. Everything gets coalesced into planes, then megaplanes, all the way up to a “hyperpage” (4,608 planes!). If you get the programming model right, and line things up, long story short…you can use it as a KV-cache and it works pretty well. They basically implemented (partial) concurrency from scratch, which is pretty neat because it’s a very OS-architecture style approach. If you understand *any* of this stuff, I’d say you have about a ~3 year head start on the future. TL;DR Raw HBF is going to be ugly, the first “somewhat useful” HBF will probably be kinda FLINT-like (their approach is fairly cheap hw-wise), and longer term I expect FlashAccel type designs to be the “actually useful” tier.

It took one week for massive parallel Sol autoresearch to find the key idea. But even more impressive is that it took the authors only one week to clean it up, improve it and turn it into a high quality paper! @kaizhengcs very curious how your week went!

People have talked about making client-side proving a reality for years... ...but now a versatile future-proof production-ready system is finally available! Another great step moving ZK from theory to practice

153 million peoples private details are for sale on the dark web. That's half of the US! All because one KYC company messed up. This sadly happens regularly. Data harvesting KYC needs to stop NOW. No-one should collect such information. What drove me to build ProveKit is showing the world that this is not necessary. Its released today. Open source. Free. Audited. Production ready. The benchmarks and examples show you exactly how to do a age check with perfect privacy. Next versions will get only better. krebsonsecurity.com/2026/09/…

We have been working very hard over the last two years to build ProveKit! A new zero knowledge proving toolkit built on top of Noir. It already powers the World ID protocol and makes it perform well on low end devices, is post quantum secure and fast! world.org/blog/engineering/p…

Will be going on @MTSlive soon to talk about the provekit.org v1 announcement and how zero knowledge cryptography will reshape privacy, digital identity and World in the age of AI.

@worldcoinfnd's ProveKit brings zero-knowledge proofs, critical, open-sourced privacy technology, mainstream when consumers and enterprises need it more than ever.