prover space rn

technical update: we have completely retired reliance on general dstorage solutions(swarm, ...) with direct p2p proof transfer channels. we expect to prove our very first eth block in the next few weeks. meanwhile, we are working on a proof-friendly dstorage layer.

we really like how RISC Zero team handled resolving (keccak) assumptions with unions. previously, one had to order assumptions and resolve them "sequentially". this was specially trouble-some in p2p proving settings where proofs could arrive in any order. nice job team 🍓

"gigagas rtp" baby that's the tweet!

E T H E R E U M

ZK all the things.

let's scale!

Proofs, not promises.

embrace succinctness anon

happy 10th birthday Ethereum 🎂

10 years of Ethereum.

Yesterday Ethereum turned 10. Today, lean Ethereum is unveiled as a vision—and personal mission—for the next 10 years. We stand at the dawn of a new era. Millions of TPS. Quantum adversaries. How does Ethereum marry extreme performance with uncompromising security and

each evm gas translates to ~20 risc-v cycles when proving eth L1 blocks. a 100m block, projected for early 2026, would thus require 2b cycles. even with with 1ghz execution rate, it would take 2s just to generate traces. "subblocks" is the way to go.

♢

good idea to run rsp on r0 😎

ai needs succinct verification for safety guarantees and that's a one-way relation. it is up to zk teams to either save cycles and prove only what matters or chase every flop.

zk everything

1/ Jolt just got a 6× speedup. We now prove over 1,000,000 RISC-V cycles/sec on a 32-core CPU and over 500,000 cycles/sec on a MacBook, with proofs of about 50 KB. Here’s what’s new 🧵

The Ethereum Foundation is reportedly offering a $1M bounty for proving key ZK conjectures. For context: Very few problems in mathematics carry $1M bounties: most famously the Millennium Prize Problems (P vs NP, Riemann Hypothesis, etc.) Ethereum is going ALL IN on ZK.

technical update: we've ported Succinct's zkevm rsp-subblock to RISC Zero. caveats: - the subblock guest is not working yet. - the unified guest(no block division) is still inefficient and needs 5x more cycles compared to when executed within sp1. github.com/WholesumNet/rs…