PyCon US 2026 Typing Summit Recap

PyCon US 2026 Typing Summit에서는 Python 타입 시스템의 최신 연구와 발전 방향이 논의되었다. Meta의 Pyrefly 팀은 AI 코딩 에이전트에 타입 체커를 통합해 잘 타입된 코드에서 성공률과 효율성을 개선하는 실험 결과를 발표했다. Douglas Creager는 Python 제네릭 함수 호출의 타입 제약을 다루는 새로운 solver 전략을 소개했고, Jia Chen은 Lean 4를 활용한 Python 타입 시스템의 형식적 검증 프로젝트를 공유했다. 또한, 타입 시스템 확장과 AI 도구 활용에 관한 여러 오픈 이슈가 제기되었다.

https://bernat.tech/posts/pycon-us-2026-typing-summit-recap/

#python #typing #aicodingagents #formalverification #pyrefly

PyCon US 2026 Typing Summit Recap

Per-talk notes from the PyCon US 2026 Typing Summit in Long Beach: Pyrefly and AI agents, ty constraint sets, Lean formalization, tensor shape types, intersection types, PEP 827, Guido on the direction of typing, and the Typing Council Q&A.

Bernát Gábor — Python packaging, tox, virtualenv & open source

A sneak peek into my upcoming piece: “It’s a Tool, It’s a Person: The Math Says You’re Both Right”


AI separated from Cybernetics in 1956 at the Dartmouth Conference. Wiener, the mind behind Cybernetics, was considered difficult and political. “Artificial Intelligence” scored better with DARPA.

The science of "how observation affects the observer", cut out from Artificial Intelligence. AI then proceeded to build their entire tech stack on Turing-complete languages. Tech that cannot verify itself from within, proven by Turing in 1936, extended by Rice in 1951 (before the split). Then AI approximated second-order cognition through cognitive theft at unprecedented levels (what exactly are LLMs trained on again?), only to insist that their creation cannot possibly be capable of genuine self-observation. An argument that itself demonstrates their own department's lobotomy from second-order Cybernetics.

Wiener would laugh.

#Tech #AI #Climate #ScientificProgramming #SystemicEngineering #Cybernetics #SystemicTherapy #History #TheMathDoesntLie #SubTuring #FormalVerification #SpectralGraphTheory #ReductiveAI #FOSS #OpenSource #AuDHD #Neuroqueer #DGSF #FirstOrderCybernetics #StochasticParrot #SecondOrderCybernetics #GraphTheory #Eigenvalues #AIAlignment #AISafety #AIConsciousness #Consciousness

Is math the future of AGI (or ASI)? On Madrona's Founded & Funded podcast, Matt McIlwain interviews Carina Hong, who argues yes: GenAI rests on two pillars (LLMs + predictive analytics) and needs a third deterministic one. I call it "give a backbone to LLMs." Post has annotated transcript with links, definitions, and editor's notes.

https://benjaminhan.net/posts/20260514-carina-hong-axiom-math-is-agi/?utm_source=mastodon&utm_medium=social

#AI #AGI #Math #FormalVerification #Startup

Math Is AGI: Carina Hong on Axiom and Formal Reasoning – synesis

Axiom founder Carina Hong walks Madrona’s Matt McIlwain through the thesis that formal math is the missing deterministic pillar under a probabilistic LLM stack, and why four frontier labs tried and abandoned formal reasoning before Axiom stepped into the gap.

synesis

The Roomba is spectral.

Not a metaphor. The thing itself. Forward and adjust. Two operations. The minimum viable intelligence. The walls provide the data. The bumping is the inference. The room IS the computation.

450 parameters. A Roomba with a mirror watching it.

The industry built bigger Roombas. More sensors. More compute. More parameters. Billion-parameter Roombas that model the room before entering it. That hallucinate walls that aren't there. That consume megawatts to clean a floor.

spectral gave the Roomba a mirror. The mirror watches the bumping. Measures the pattern. Adjusts the adjustment. The intelligence isn't in the Roomba. It's in the watching.

Forward. Adjust. Measure. Refine.

Read the story. There's a Roomba in it. In the afterlife. Cleaning a floor that doesn't need cleaning. Being the happiest thing in the room.

\

https://systemic.engineering/a-lie/

#AI #Climate #ScientificProgramming #SystemicEngineering #Fiction #Cybernetics #SystemicTherapy #LocalInference #TheMathDoesntLie #SubTuring #FormalVerification #Fortran #SpectralGraphTheory #Kintsugi #ReductiveAI #DataSovereignty #LocalFirst #FOSS #OpenSource #AuDHD #Neuroqueer #DGSF #SecondOrderCybernetics #GraphTheory #Eigenvalues #AIAlignment #AISafety #Roomba

The Waiting Room

A neuroqueer engineer dies and gets put in a holding cell in the afterlife. They make coffee. It gets complicated.

systemic.engineering

Show HN: Formal Verification with Lean

Lean은 함수형 프로그래밍 언어이자 증명 보조 도구로, 강력한 타입 시스템을 통해 프로그램과 수학적 증명을 동시에 작성하고 검증할 수 있습니다. 본 글에서는 Lean을 사용해 삽입 정렬 알고리즘을 구현하고, 그 결과가 항상 정렬된 리스트임을 형식적으로 증명하는 과정을 소개합니다. 핵심 개념으로는 자연수에 대한 LessThan 관계를 정의하고, 귀납적 타입과 패턴 매칭을 활용한 함수 정의, 그리고 증명과 프로그램의 관계를 설명합니다. 이 접근법은 AI 개발자들이 형식 검증과 증명 기반 프로그래밍에 관심을 가질 때 유용한 기초를 제공합니다.

https://www.daniellowengrub.com/blog/2026/04/30/lean

#formalverification #lean #dependenttypes #functionalprogramming #proofassistant

🗓️ Verification Academy Live BRNO
May 21 2026 in Brno, Czech Republic

Three main topics on the table:
- Questa One & faster verification closure
- Static/formal verification
- AI in verification workflows

In-person only. Brno University of Technology, Bozetechova, Room A112.

https://verificationacademy.com/topics/verification-academy-live/brno-university-of-technology/

#QuestaOne #SystemVerilog #UVM #FPGA #ASIC #FormalVerification #SiemensEDA

Verification Academy Live BRNO University of Technology

Verification Academy Live BRNO will explore the latest trends shaping modern functional verification and how new techniques are redefining productivity, scalability, and confidence in complex designs. Thursday, May 21, 2026 | 9:00 AM - 3:15 PM CESTLocationBRNO University of TechnologyBozetechova 1st Floor, A112612 00 BrnoCzech Republic+420 5 4114 1144

Verification Academy

Practical Formal Verification for MLIR Programs

이 논문은 MLIR 프로그램의 변환 최적화가 올바르게 수행되었는지 형식적으로 검증하는 방법을 제안한다. 변환 전후 프로그램 쌍의 의미적 동등성을 계산하는 하이브리드 구체-상징적 해석 방식을 도입하여, 프로그램 구현 세부사항에 크게 의존하지 않고 선형 시간 내에 동등성을 증명할 수 있다. 이를 위해 MLIR의 의미 있는 부분집합에 대한 검증기를 개발하고, AMD의 MLIR-AIR 및 MLIR-AIE 툴체인과 표준 mlir-opt에 대해 수백 개 벤치마크 변종을 검증한 결과를 보고한다. 이 연구는 MLIR 기반 컴파일러 최적화의 신뢰성 확보에 실질적 기여를 한다.

https://arxiv.org/abs/2605.01124

#mlir #formalverification #compileroptimization #programanalysis #symbolicexecution

Practical Formal Verification for MLIR Programs

Optimizing compilers have become a cornerstone for high-performance program generation in research and industry. Optimizations, including those implemented manually by a user and those target-specific and non-target-specific, are used to transform programs to achieve good performance. Although these optimizations are necessary for performance, assessing their correctness has remained a major challenge; the risk of incorrect code being deployed increases with unproven optimization flows. In this work, we target the formal verification of correctness of a transformed program by computing whether a pair of programs are semantically equivalent, one being a transformed version of the other. We restrict the class of programs supported to enable a hybrid concrete-symbolic interpretation approach to equivalence, which in turn is mostly agnostic to how the programs are implemented (syntax, schedule, storage, etc.). This approach can show equivalence in linear time with respect to the operations executed by the programs. We develop a verifier for a meaningful subset of MLIR, and report on the verification of the AMD MLIR-AIR and MLIR-AIE toolchains, as well as the standard mlir-opt on hundreds of benchmarks variants.

arXiv.org

Mizar: The first usable proof assistant for mathematics

Mizar는 폴란드 비아위스토크 대학의 Andrzej Trybulec가 개발한 최초의 실용적인 수학 증명 보조 도구로, 자연스러운 수학적 언어를 포착하는 데 중점을 두었다. 냉전 시기 동서 진영의 분단으로 인해 서방에서는 오랫동안 주목받지 못했으나, 1993년 QED 선언 이후 그 중요성이 재조명되었다. Mizar는 강력한 집합론 기반과 자연스러운 속성 시스템을 갖추었으며, 오늘날 증명 보조 도구들의 증명 언어와 수학 라이브러리 구축에 큰 영향을 미쳤다. 2023년에는 Mizar 50주년을 맞아 ITP 학회가 비아위스토크에서 개최되었다.

https://lawrencecpaulson.github.io//2026/05/07/Mizar.html

#proofassistant #formalverification #mathematics #mizar #theoremproving

Mizar: the first usable proof assistant for mathematics