testing ethereum | dabbles in anxiety | 🏴🇪🇺/acc
small fork energy
Joined October 2022
- Tweets81
- Following272
- Followers100
- Likes284
spencer 🏴 retweeted
Brother get off your high horse. The world is more than NYC.
I bet you’ve never even been to India. There’s a thriving Ethereum community there and has the most rapid growth of crypto devs.
All you guys bitching and whining. Super lame.
Not sure what they expected hosting it in Mumbai, generational L.
Pretty wild though, this is the first year in crypto I can remember where there wasn't a major Ethereum based conference everyone went to.
Obviously I'm biased here, but I think within 1-2 years the only major "crypto" conferences will be DAS and Token2049.
We're in a world where no one cares what chain it's built on, the ecosystem conferences of the past will get smaller and smaller.
On a final note, I think crypto companies are sleeping on tech and SF based conferences.
It's well understood that crypto is merging with finance, but there's also a lot of overlap with AI.
Big opportunity for enterprising teams who can get to this market first.
spencer 🏴 retweeted
worth to also have a look at:
Most SSZ libraries are tested against the spec. I built one where the core properties are formally proved, in Lean 4.
It also passes the full upstream test corpus. SSZ is how Ethereum serializes and Merkleizes its data — if hash_tree_root or deserialize is wrong, you get the kind of bug that can split the chain. For this reason I wanted something that is not only tested against the spec, but also proved.
Formal verification. The library is called SizzLean. I proved three theorems: roundtrip (decode after encode returns the same value), non-malleability (two different values cannot produce the same encoding), and a size bound. They hold for the fixed-size SSZ types: the integers, bool, fixed vectors and lists, and containers built from those. I have not proved them yet for bitvector, bitlist, and variable-size containers. So this is a base, not a finished proof.
Trust boundary. The only thing I assume without proof is the C SHA-256 function, declared as three named Lean axioms. You can list the full trust footprint with one grep, or with #print axioms on any theorem. Nothing else is taken on trust.
Two backends. The same spec runs on a fast cached backend, used for execution and for running the test vectors, and on a pure backend, used for the proofs. I put the hash function behind a typeclass, so SHA-256 can be replaced by Poseidon2 or a post-quantum hash without changing the containers, the proofs, or the cache.
Conformance. On top of the proofs, it runs the official ethereum/consensus-spec-tests on both presets:
ssz_generic: 2188 / 2188
ssz_static (mainnet): 1585 / 1585
ssz_static (minimal): 38991 / 38991
This covers every fork from Phase 0 to Gloas, including the new ePBS containers.
I also commented the code to be read. If you want to learn how SSZ works, or how to do this kind of work in Lean 4, you can follow it from top to bottom.
The verification was done with heavy help from an AI coding agent. With Lean this does not change the result: the kernel checks the proof, independently of who wrote it.
I work on specs and test vectors at the Ethereum Foundation, on the STEEL team. This is a personal project. The next step is to widen the proofs toward the remaining types.
Repo and full details: github.com/etheorem/etheorem…
spencer 🏴 retweeted
Last week, core devs had to delay forking the Hoodi testnet to Glamsterdam to study a report by @chain_security which revealed some problems with how EIP-8037 interacts with a few contracts.
Over the weekend, ppl tried to get a better picture of the impact and by now it's quite well understood. So here's what you need to know:
What's been known is that contracts which hardcode forwarded gas are affected by repricings, e.g. making state writes (SSTORE) more expensive.
What's new is that some coding patterns rely on specific invariant which were broken with EIP-8037.
Gas already spent can come back mid-tx, so gasleft() can go up. This causes two problems:
(1) a before / after gas counter underflows and reverts
(2) a failed sub-call hands gas back, so a path that should've been unreachable runs, and the tx looks like it succeeded
Related but different: (3) measuring the gas usage of a call using GAS - before and after the call - might undercount because the state gas reservoir is not visible during execution (e.g. gas_left).
It's now about deciding how to proceed. Delaying hoodi bought time but won't resolve the issue. This week, we have to dig deep into contracts and make sure nothing is missed. Eventually, this might just need some coordination among the involved parties (dapps, core devs, auditors, people running nodes, etc.), or worst-case, spec changes.
If you're an app dev or auditor, check your contracts for issues with EIP-8037/EIP-8038, and report them.
Also, share this with your favorite app dev to spread awareness.
⚠️ It has just been decided to delay the Glamsterdam hard fork on the Hoodi testnet because of security impacts we identified in this blog post: chainsecurity.com/blog/glams…
We already found 13 real-world examples and are in responsible disclosure with the affected projects.
Please check your contracts as soon as possible, as the EF is determining how many contracts are impacted in which way.
Feel free to reach out to us with questions via [email protected].
spencer 🏴 retweeted
After 3h and:
- Not a single payload missed
- No issues
My @go_ethereum impl fully self-migrated from the MPT to the Partitioned Binary Tree with
- 2TB disk
- 16GB of RAM!
Roots matching🎉
I'm very happy with the results and proud of this last months of hard work!
⏰ I* ⏰
It's happening!🚨🚨🚨
The next @ethereum's tree Partitioned Binary Tree is finally here! And we're migrating to it!
We are doing the first tree & DB migration in the Glamsterdam-devnet-8 with @go_ethereum @Nethermind @Besu_eth @ErigonEth
This is gonna be huge! Ready for I*🎉
Tradfi 2026 Q2 stats as a thought experiment :D
L1s
- Fedwire (US) $304T
- T2 (EU) ~$143T
- CHAPS (UK) ~$34T
- Ethereum $7.1T
L2s (settle on an L1)
- CHIPS ~$125T
- ACH $25.9T
- Ethereum L2s $9.65T
L3s (settle via banks)
- Visa $4.0T
- Mastercard $2.4T
spencer 🏴 retweeted
Today the EF and OA announce zkAPI—a new means for private AI
Picture this: deposit ether into a vault, sign a zk proof, get a fixed amount of private inference. Advances like these help us move past middlemen's requirements that we link our data and identities.
Read more below
Quick updates on the Open Anonymity project (@OpenAnonProject)!
1. OA-chat is live! You can now signup: chat.openanonymity.ai
2. We collaborated with @ethereumfndn (@VitalikButerin, @DavideCrapis, @VittoStack, @curryrasul) to develop zkAPI, a new zero-knowledge protocol for accessing gated / paid APIs with unlinkability. It’s like a generalization of OA unlinkable inference that also abstracts payments away. Now also available on OA-chat!
I'm excited because more infrastructure should exist where privacy and sovereignty is foundational and people are in control.
spencer 🏴 retweeted
Security researchers, get ready! The Glamsterdam upgrade is approaching. The Ethereum Foundation’s bug bounty program will begin accepting reports for Glamsterdam bugs once the Sepolia upgrade blog post is published. And take note of the reward multipliers!
ethereum.org/bug-bounty/#gla…
spencer 🏴 retweeted
I just read my first AI book.
Two months ago, I asked ChatGPT to imagine what the world might look like in 2050.
The answer was very interesting, so I asked it to write an entire fiction book set in that future.
It generated a ~100-page story about a world where AI gets so good at predicting human behavior that it can prevent tragedies before they happen.
The story explores a simple question: if AI could make the world safer by slowly taking away our ability to make mistakes, would we let it?
Among all the futuristic visions in movies and books, this one feels the most plausible to me.
The v1 of the book was a bit slow, so I asked ChatGPT to make a v2 more dynamic.
Here’s the book: drive.google.com/drive/u/0/f…
The chat expired, so ChatGPT made a new version. We might not be reading the same version.
spencer 🏴 retweeted
my bend evm seems to pass all the Amsterdam Ethereum HF state tests - its extemely slow but its worth taking a look as i diff tested w revm/evm2 and looks good
github.com/gakonst/evm-bend
i have a strong feeling jev (or openai/anthropics future equivalent) will be integrated heavily into existing llms -> codex/cc -> significantly reduced token costs… the future is exciting
spencer 🏴 retweeted
We support EIP-8198 (Quick Slots) and @ethlabs_org’s push for faster Ethereum blocks!
Great read on the trading costs of slow blocks. Our own @josh_pwrk explains how arbitrageurs could take less price risk while waiting for their L1 spot trades to land, letting them step in on smaller price gaps and help keep spreads tight.
We're bringing Derive V3 to Ethereum L1, so we’d love to see this ship.
spencer 🏴 retweeted
🎄🌲When Binary Tries?🌲🎄
You're witnessing @Nethermind @ErigonEth @go_ethereum and @Consensys 's Besu all handling a reorg while moving into Partitioned Binary Trie (eips.ethereum.org/EIPS/eip-8…)
You can't get it more solid than this!
Next step 👉 Mainnet migration tests
👇👇
🏴 🇪🇺 ... make it happen and i will move back
as open source ai continues to follow the frontier models, im going to assume the cost of agentic cyberattacks gets very cheap, in the order of millions of dollars
for me p(humanity wiping out humanity) > p(ai on its own wiping humanity)
alas its a very complicated and nuanced problem and i know nothing