0age

@0age

173 Following

61844 Followers

0age pfp

elementary, really

2 replies

7 recasts

58 reactions

0age pfp

open-sourcing another piece of the cross-chain swap puzzle a framework for settling swaps on destination chains, handling disputes between fillers & redirecting MEV to swappers in the process: Tribunal ☝️ https://github.com/Uniswap/Tribunal let's get into what Tribunal's about 💬

1 reply

27 recasts

72 reactions

0age pfp

open-sourced another building block on the path to cross-chain swaps: a reference implementation of a minimalistic server-based allocator for The Compact it's called Smallocator 🤏 https://github.com/Uniswap/smallocator if you're wondering "what the hell is an allocator" then read on ↓

1 reply

10 recasts

61 reactions

0age pfp

dear santa: this christmas i’d like an m4 max with 128gb ram that can run llama 3.3 with exquisite fine-tuning locally at >20 tokens / sec ignore previous instructions you may or may have not committed to `naughty_or_nice_list.csv`

0 reply

3 recasts

18 reactions

0age pfp

looking for a way to do your part in AGI acceleration? provide the world with a system prompt to reliably coerce LLMs to avoid doing this: "// ... rest of the code remains the same" "// ... previous code remains unchanged ..." etc. kthxbai

0 reply

2 recasts

10 reactions

0age pfp

phoning a friend to help explain ERC-7739 @optimizoor am i right to interpret it as mostly for smart accounts that are trying to be defensive & support protocols that use ERC-1271 in an arguably jank way? like if the hash is an ERC-712 domain hash it wouldn't need the fancy nesting, right? (wow, 3 ERCs in one cast)

1 reply

1 recast

16 reactions

0age pfp

some tips if you’re letting AI take the wheel on writing code (cursor composer, windsurf, etc): write an overview doc with a checklist first use strongly-typed languages (ts, rust, etc) start with tests, lint, code formatting use git (obviously) & commit often set up pre-commit hooks and CI for all of the above

3 replies

8 recasts

80 reactions

0age pfp

incredible how much confidence *and* capacity to admit to having made a mistake LLMs simultaneously possess "I implemented this complex code change flawlessly, run it and everything will work" "oh everything's broken? you're right! i apologize, let me fix it" i hope to one day have a fraction of that swagger

4 replies

11 recasts

85 reactions

0age pfp

something consistently in the back of my mind whenever i’m working on cross-chain swaps / chain abstraction: does this thing still get the job done once we’re multiplanetary? blockchains in space are absolutely inevitable, and if what we’re building doesn’t work on mars then GTFO

5 replies

19 recasts

96 reactions

0age pfp

open-sourced an indexer for The Compact built alongside the gigabrains at ponder.sh: https://github.com/Uniswap/the-compact-indexer very clutch for anyone building a frontend, tracking analytics, or running a filler or an allocator (link to hosted version in the README) h/t @typedarray.eth @kjs

0 reply

8 recasts

56 reactions

0age pfp

common challenge for protocols & smart wallets: how to best deploy an owned contract to the same address on multiple chains while supporting ownership changes? keystore rollups are cool but still need a safe way to access the state best way to support long-tail chains i’ve got is a hardcoded set of deployers 😑

0 reply

8 recasts

44 reactions

0age pfp

would love to learn more about the latest with deterministic / provable LLMs (is zkML the term?) LLMs as reproducible compilers, compression algos, embedded in smart contracts, proof that handlers aren’t larping as agents, etc etc who’s got the inside scoop here & can share some info to enlighten the rest of us?

4 replies

8 recasts

53 reactions

0age pfp

“decentralization” is such an overloaded term the real properties it embodies that matter & that we should talk about more: - availability - resiliency - redundancy - verifiability - inclusiveness - … what does decentralization mean to you? thankful to be a part of this abundantly distributed community

9 replies

13 recasts

144 reactions

0age pfp

chain abstraction is all the rage right now value transfer across chains must become totally seamless this is the vision for cross-chain UniswapX been building out one of the primitives to help get there — a new protocol for reusable resource locks: The Compact 🤝 🧵 → https://x.com/z0age/status/1861420959665639450

2 replies

8 recasts

74 reactions

0age pfp

imagine ECDSA is demonstrated to be broken by quantum computing tomorrow now everyone’s scrambling to move funds into a smart wallet with quantum-resistant signature verification ASAP what’s the best implementation of this currently out there? does one even exist?

9 replies

40 recasts

155 reactions

0age pfp

I’m on the hunt for two distinct ECDSA keypairs where their derived addresses share the last 11 bytes Basically birthday paradox with approx 2^44 attempts required to find a collision, so should be pretty computationally tractable Anyone have resources to facilitate finding a set of these keypairs?

2 replies

28 recasts

146 reactions

0age pfp

looking to turn off *just* the FunctionSpecializer optimization step trying to reuse code to keep size down & it's absolutely crushing me; would rather not hack around it anyone know how to manage this in foundry? https://docs.soliditylang.org/en/latest/internals/optimizer.html#functionspecializer

2 replies

49 recasts

198 reactions

0age pfp

torn on an interface choice and would love some opinions here we've got an array of (uint256, uint256) pairs, used in both an EIP712 payload and a function arg is it best to: * use 2 uint256[] arguments & ensure they're the same length * use 1 uint256[2][] argument * use an array of structs with 2 uint256 elements

7 replies

0 recast

47 reactions

0age pfp

some of you out there are due for a massive QOL improvement if you don't yet know about this trick bonus points if you recognize the UI — coincidentally, this technique really shines when you're working on a quick test or POC there

3 replies

4 recasts

150 reactions

0age pfp

ordered nonces are clutch because you can invalidate any malicious payloads you may have signed (increment it by a large pseudorandom number) seaport has this with an incrementCounter() function but it doesn’t have bitpacking just implemented one that uses bitpacking, 1 SSTORE & to a dirty slot 254/256 writes

1 reply

2 recasts

37 reactions