Functional Algorithms, Verified

0 comments

Lobsters
Formal Methods

0 comments

Lobsters

Lean 4.29.0 released with 453 changes.

Highlights: reduced startup time through static initialization of closed terms, simpler ๐š—๐š˜๐š—๐šŒ๐š˜๐š–๐š™๐šž๐š๐šŠ๐š‹๐š•๐šŽ semantics improving predictability, higher-order Miller pattern support in ๐š๐š›๐š’๐š—๐š's e-matching engine, and a significant overhaul to instance and reducibility handling.

Full release notes: https://lean-lang.org/doc/reference/latest/releases/v4.29.0/

#LeanLang #LeanProver #FormalMethods

Validating Hareโ€™s Sort Module using Symbolic Execution via @RunxiYu https://lobste.rs/s/yoq0e6 #formalmethods #plt
https://notes.8pit.net/notes/y7n8.html
Validating Hareโ€™s Sort Module using Symbolic Execution

0 comments

Lobsters
Red-black tree in Lean 4 prover with everything proved https://lobste.rs/s/zd797w #formalmethods #plt
https://rentry.co/8sfon8ez
Red-black tree in Lean 4 prover with everything proved

0 comments

Lobsters

If your security depends on someone reacting in time,
you donโ€™t have a guarantee. You have coordination.

Wrote about a real failure mode that keeps showing up.
https://app.daily.dev/posts/yqybiiFvJ

#blockchain #smartcontracts #security #formalmethods #web3 #cryptography #defi #ethereum #protocoldesign

On-Chain Finality and Off-Chain Assumptions: A Failure Mode in Dispute Resolution Systems | daily.dev

Blockchain dispute resolution systems often present themselves as deterministic, but many rely on a hybrid model combining on-chain state machines with...

daily.dev

fly51fly (@fly51fly)

๋Œ€๊ทœ๋ชจ ์–ธ์–ด๋ชจ๋ธ์„ ํ™œ์šฉํ•œ ํ˜•์‹์  ๋ฐ˜๋ก€ ์ƒ์„ฑ ์—ฐ๊ตฌ 'Learning to Disprove: Formal Counterexample Generation with Large Language Models'๊ฐ€ ์†Œ๊ฐœ๋˜์—ˆ๋‹ค. ์ •๋ฆฌยท๊ฒ€์ฆ์ด ํ•„์š”ํ•œ ์ˆ˜ํ•™/๋…ผ๋ฆฌ ๋ฌธ์ œ์—์„œ LLM์„ ๋ฐ˜๋ก€ ์ƒ์„ฑ ๋„๊ตฌ๋กœ ์“ฐ๋Š” ์ƒˆ๋กœ์šด ์—ฐ๊ตฌ ๋ฐฉํ–ฅ์„ ์ œ์‹œํ•œ๋‹ค.

https://x.com/fly51fly/status/2036199167882830168

#llm #research #formalmethods #counterexample #arxiv

fly51fly (@fly51fly) on X

[AI] Learning to Disprove: Formal Counterexample Generation with Large Language Models Z Li, Z Li, K Yang, X Maโ€ฆ [ETH Zurich & University of Toronto & MiroMind] (2026) https://t.co/aqE7QU8npQ

X (formerly Twitter)
TLA+ mental models

0 comments

Lobsters