Parker Thompson retweeted
Our open models are here.
Both of them.
openai.com/open-models
One of the coolest parts of this is 'pipelining' which allows us to schedule the prove tasks while the executor is still running. So with ideal saturation of GPUs to execute speed our e2e proving time is only a few % less than execution speed.
(Scaling this workflow is a fun job)
Linear is so frustratingly close to being good with GitHub but really falls short on a couple key integration points that force extra process and overhead onto devs that doesn't need to exist imho
Parker Thompson retweeted
Introducing zkVM 1.0: The Future of Verifiable Computation🌐
Today, we're excited to launch the world's first and most performant, production-ready zkVM.
zkVM 1.0 makes zero-knowledge proofs scalable, accessible, and cost-effective for any chain.
All engineering managers should prepare for mysterious principle eng pto
CIVILIZATION VII. Coming 2025.
Wishlist now & watch the full reveal this August!
civilization.2k.com/seven | #Civ7 #SummerGameFest
Parker Thompson retweeted
One of my earliest claims to fame was actually in cryptography.
Remember WEP Wi-Fi security?
Back in the day I wrote this software called AirSnort that broke that encryption.
It got me my first interview ever in Wired in 2001:
wired.com/2001/08/wireless-n…
Parker Thompson retweeted
Our zkVM Performance Upgrade Roadmap for Q1, Q2 2024 is live. TLDR:
✔️ 7-10x improvement in global performance
✔️ Upgrades and new features for custom accelerator circuits
✔️ 20x + cost improvements by the end of 2024
Learn more: r0.link/perfupgrades
Parker Thompson retweeted
Our latest demo, built using the RISC Zero zkVM, unlocks identity for Web3:
✅ Prove & mint your identity as a soulbound NFT
✅ Maintain on-chain privacy, stay compliant
✅ Links Web2 authentication with Web3 for easy implementation
Start building ▶️ r0.link/zkKYC
Parker Thompson retweeted
Welcome @SuccinctLabs 👋 and congrats on SP1, an impressive accomplishment and a great signal for the future of general purpose zkVMs.
We’re focused on providing the cheapest, most scalable, and most effective platform for building and deploying ZK.
Since benchmarks against our tech were referenced, we want to provide some clarification:
- We’ve achieved SP1 equivalent performance using GPUs.
- We’re optimized for cheaper, more readily available hardware – 6 times cheaper than the SP1 rig. We haven’t spent effort optimizing for CPUs.
- Comparing $/megacycle we’re 75% of SP1, and that number is falling fast. High speed is good, low TCO is better.
- We’re fully committed to open source with our circuit construction & composition system (i.e. precompiles) still under heavy internal development, but coming soon.
- We have supported std and no_std for a long time (cc: @gakonst)
More specifics next week…
If these are the people we hoped would self-regulate AI, law makers should have stepped in a while ago.
This has been a really fun exploration of how we could leverage oidc identities in non standard environments like Ethereum.
Plus check out the cool trick of how we stuff data into the nonce ;)

















