DeFi is crippled by static parameters—swap fees, interest rates, and risk thresholds lag markets by months. @Balancer is integrating @ezklxyz to pioneer dynamic fees that adapt to market volatility in real time, trustlessly. 🧵 No more static parameters—here’s how it works and why it matters.

ZKONNECT is coming to @EFDevcon. We’re bringing together the builders pushing ZK from research to production, where zkVMs, ZK Data Coprocessors, zkML, and zkTLS meet real-world deployments and scale. Featuring a special speaker from @eth_proofs, co-hosted with @HouseofZK, and supported by an incredible lineup of partners making this possible. Speakers and panels to be announced soon. 👉 Save your spot: luma.com/jix04ltq

ZK makes your infrastructure idiot-proof. Even if someone gains root access to your server, they can kill it but can’t generate malicious commands that trick downstream verifiers controlling critical systems (water management, oil rigs, defi protocols etc…).

Have been really interested in silent data corruption at scale lately and @joakimhi sent me this paper from Villanova/Meta that uses a simple statistical analysis of neuron activation patterns to detect and fix these silent errors before they cause system failures.

There's massive latent demand for advanced cryptography (ZK, FHE, ..) outside of crypto. But most companies building this tech are funded by crypto VCs or foundations, so they default to serving blockchains. Meanwhile Google is lapping all of us.

Hey Sean Cook if you want to make sure you can KYC your users without ever ever having to see their data or store it on the cloud check out @ezklxyz and @sundialxyz for privacy preserving and secure ID solutions :)

To get back into GPU programming , I read @Si_Boehm's amazing primer on how to optimize CUDA matmul kernels. Given most of large network FLOPs are from matmul its a critical set of optimizations to know (lots of deepseek's opts for eg. came from an optimized matmul kernel).

Pretty interesting finding: Water-cooled datacenters are at much higher risk of bit flips than other types of cooling. Research from NTT & Hokkaido shows hydrogen in water thermalizes cosmic ray neutrons into the exact energy range that causes the most soft errors.

Prod AI systems face a fundamental reality: they’ll eventually encounter the worst possible conditions. Hardware failures, adversarial attacks, cosmic rays flipping bits in memory. If it can break your system, it probably will. We’re hiring a DevOps Engineer who gets this. 🧵

1/x Underwriting in DeFi is broken - centralised, opaque, and trust-based. Qiro is changing that with verifiable credit risk assessments powered by @ezklxyz Here’s how we’re levelling up underwriting in DeFi 👇 qiro.substack.com/p/qiro-x-e…

We ran into this exact issue when building our cluster. Servers had defects and high temps would cause memory corruptions. But a really cool byproduct of software-base verifiable compute like @ezklxyz is that you catch these failures immediately. A single bit flip and you get a computation that does not verify and triggers alerts.

EZKL’s mission to make AI verifiable by default is starting to resonate very strongly outside of crypto. As AI is inserted into critical paths; creating PRs, managing energy pipelines, accessing sensitive data; security is top priority. Stay tuned for coming updates 😘

A glimpse into @ZuBerlinCity cryptography agenda: the snapshot of the most hot and sharp topics inhabiting blockchain --> 'risc v or not risc v' a panel with @leonardoalt from @powdr_labs, @levs57 (online) from @alloc_init_, @carlo_cretino50 from @0xLita and more --> 'FHE without overhead, 6x improvement to Zama and Duality performance' by Ilya Rososhek from Mathlock --> 'Is folding dead?' with @liameagen from @AlpenLabs --> 'Google's zkID: is it zk adoption we hoped for so long?' a panel with @therealyingtong from Ethereum Foundation, @madztheo from @ZKPassport and @olkurbatov from @Rarimo_protocol --> '25 zkVMs: Too few or too many?' a panel with @ETazou from @KakarotZkEvm, @jimpo_potamus from @IrreducibleHW, and @Zac_Aztec from @aztecnetwork --> 'co-snarks: technology evolution and where we are heading to' by Franco Nieddu from @TACEO_IO --> 'Can we ever zk-prove LLMs' by @CamutoDante from @ezklxyz --> 'iO: hopes and doubts' by @backaes from @PrivacyScaling @machina__io --> 'ZK-VM fuzzing' by @codytouchgrass from Ethereum Foundation --> 'Lattice-based snarks' by @KhanhCrypto from @KingsCollegeLon --> 'zkTLS: is it real?' a panel by @sylvechv from @hyli_org, @Mojtek from @vlayer_xyz and more --> 'ZK formal verification' by @hackerdocc --> 'Emulating OP RAND in Bitcoin' by @olkurbatov --> 'zkLinux' by @stskeeps from Zippie --> 'Deep dive into FFTs' by Matan Prasma from Ethereum Foundation And your bonus sessions: --> 'How to stay sane in crypto' a fireside chat by @hudsonjameson and me --> 'Technical education in 2025: tech writing, twitter, tiktok and what not' by @DAppaDanDev and @portport255 from @zksync Several more talks TBA (Poseidon hash, state of quantum cryptography, state of art ZK, real world crypto, etc) The last but not the least: it is still one month till ZuBerlin happens – you have enough time to change the plans :P Apply here >>> join.zuberlin.city/apply

Corporate interests have hijacked much of AI research. When I did my PhD at Oxford in 2018, it was clear how deeply the field was shaped by the priorities of FAANG giants that funded our research.

Ever since I read this paper I've been thinking about it, lately I was dreaming about Asimov's "I, Robot" and the thought occurred to me, can we use ZKPs to verify the three laws of robotics? So I wrote a note about it, if you're interested, check it out! ziemann.me/ai-crypto

this is a privacy and verifiability bull post i write about privacy and verifiability tech as a historically underperforming venture vertical and why this time may actually be different given the catalyst of accelerating AI proliferation in our everyday lives i have increasingly found myself defending and evangelizing @worldcoin (a vast departure from what my preconceived notions of world were one year ago) so i decided to write about it i talk through adoption of zk @GooglePay @Cloudflare zktls @digg @reclaimprotocol zkml @Balancer @ezklxyz i also touch on experiments with TEEs and FHE in crypto @CatalystLabsX @chainpro_xyz @inconetwork

Zero knowledge proofs are going mainstream and being integrated into Google products

played with tees, and I am now way more appreciative of zk/fhe. sorry for straying 😭 software >>>>> hardware in long run™️. It can takes months to patch TEEs. software patches? super fast. flexibility for users? off the charts. devX with @NoirLang or @ezklxyz? chef's kiss 😘

does anybody know a really good devops person looking for a role :)?

I’ve been looking for finetuning methods that use less memory than backpropagation, and came across this super clean "forward pass only" method called MeZo. It uses less than 1/12th of the memory backprop requires by leveraging gradient-less updates.

💦💦💦