#FormalVerification
I've been trying TLA+ and Lean alongside a small Java example. One item, two callers, and a rule I want the next coding agent to find through the build. Which rule would you start with?

buff.ly/ZHlanBZ

#Java #FormalVerification
November 3, 2026 at 7:15 AM
Here's No. 5 From Proof to Binary of our 6 part vericoding overview.

This one details how we compile from proof assistant to runnable code.

#formalverification #vericoding
November 3, 2026 at 11:48 AM
Here's number 3 in our 6-part Vericoding introduction series.

Your first contract.

#formalverification #vericoding
October 30, 2026 at 1:00 PM
What Lean Verified About the 2.25 Matrix Bound

A green “kernel accepts” badge means something very specific: Lean has checked a formal proof term for a precise theorem statement.

#lean #formalverification #matrixmultiplication #theoremproving
November 3, 2026 at 10:03 PM
Continuing the 6-part Vericoding Introduction. No. 4: Two Ways to Specify.

This one looks at intrinsic and extrinsic methods and provides some handy rule of thumb guidance when performing verification for programming languages.

#formalverification #vericoding
October 31, 2026 at 12:47 PM
AI can exploit software vulnerabilities at terrifying speed.

AI-generated code and legacy systems are at risk.

At Scidonia, we use vericoding to verify that code is safe, secure and correct.

Visit scidonia.ai to learn more or book a discovery call.

#formalverification #vericoding
November 1, 2026 at 10:19 AM
No. 2 of Vericoding in 6 parts - Proofs are programs.

This video explains how and why theorem provers can be thought of as a programming language.

#formalverification #vericoding
October 29, 2026 at 12:11 PM
Extended SMT-based verification framework for quantum error-correction circuits with formal soundness proof. Verifies CSS codes, Steane code, surface codes up to distance 101, and GHZ circuits with 262,144 qubits on standard laptop.

#QuantumErrorCorrection #FormalVerification #Research
Formal Verification of Quantum Error-Correction Circuits Using Bit-Vector Abstractions and SMT Solving
iq.fp2.dev
October 31, 2026 at 10:55 AM
Extended SMT-based verification framework for quantum error-correction circuits with formal soundness proof. Verifies CSS codes, Steane code, surface codes up to distance 101, and GHZ circuits with 262,144 qubits on standard laptop.

#QuantumErrorCorrection #FormalVerification #Research
Formal Verification of Quantum Error-Correction Circuits Using Bit-Vector Abstractions and SMT Solving
iq.fp2.dev
October 31, 2026 at 10:55 AM
A Lean-Checked Proof Is Not Yet Mathematical Acceptance

Lean can establish something very specific: the formal proof file compiles, and its encoded conclusion follows from the definitions, axioms, and lemmas inside it.

#lean #formalverification #zetafunction #riemannhypothesis
October 29, 2026 at 2:01 PM
What Has to Happen Before a 7/8 Zeta Claim Is Trusted

A claimed zero-free wall is only the beginning of the mathematical process. The statement here is that ζ(s) has no zeros when the real part of s is greater than 7/8.

#riemannzeta #riemannhypothesis #numbertheory #formalverification
October 30, 2026 at 6:02 PM
Claude wrote machine-checked proofs that 239 entry points in 123 open-source libraries never crash, finding hundreds of bugs.

Start small: spec what one parser must never do, then make the agent prove it.

The Short: youtube.com/shorts/eI0l2Tdh1y8

#AI #CyberSecurity #Claude #FormalVerification
Claude Wrote Machine-Checked Proofs for 123 Libraries | 500x Engineer
Anthropic and Theorem had Claude prove 239 entry points crash-free and found hundreds of bugs. What was proven, the limits, and how to start.
www.500xengineer.com
October 28, 2026 at 10:22 PM
Just heard a rival AI startup just raised $295 M to launch an AI agent that can formally verify code—think encryption routines with provable math proofs. Could this be the next leap for software security? Dive in to see the details! #AIAgent #FormalVerification #CodeVerification

🔗
October 29, 2026 at 3:21 PM
Just heard a rival AI startup just raised $295 M to launch an AI agent that can formally verify code—think encryption routines with provable math proofs. Could this be the next leap for software security? Dive in to see the details! #AIAgent #FormalVerification #CodeVerification

🔗
October 29, 2026 at 3:21 PM
AI just cut a 5-month formal verification job to 11.4 hours for $466.99—proving crypto code correct without touching it. Big shift in who #FormalVerification #AI #SoftwareSecurity #Cryptography

https://freegardner.com/synapse/ai-automates-formal-verification-of-secure-software.html
AI Automates Formal Verification of Secure Software
AI Automates Formal Verification of Secure Software
freegardner.com
October 29, 2026 at 4:25 PM
Lean Can Check the Logic, Not Explain Why It Matters

A proof assistant can turn “assume a zero past 7/8” into a precise chain: estimate moments, apply a large-sieve bound, derive a contradiction, and rule out that zero.

#lean #formalverification #riemannzeta #numbertheory
October 28, 2026 at 6:02 PM
Delaware is drafting companies an AI agent would run, with power to own assets and sue. House Oversight wants the state's documents, and the capital floor and audit access stay unpublished. #AIAgents #FormalVerification https://sharedsapience.com/century-report/the-century-report-october-28-2026/
October 28, 2026 at 5:27 PM
Formal verification is having a moment.

This is something we strongly advocate (and provide as a service).

We've put a 6-part explainer series together explaining vericoding. Here's the first:

What is vericoding.

#vericoding #formalverification
October 27, 2026 at 10:01 AM
What Lean Actually Certified About the 7/8 Zeta Claim

A Lean check is stronger than a screenshot saying “proof complete,” but narrower than a verdict on all of mathematics.

#lean #formalverification #zetafunction #numbertheory
October 26, 2026 at 2:01 PM
OpenAIが主張する3つの科学的ブレークスルー:AI生成研究の検証プロセスと専門家査読の全貌

OpenAIが発表した数学・物理等のAI生成ブレークスルーと専門家による検証スキームを解説。

#LLM #OpenAI #AIResearch #FormalVerification #QuantumPhysics
OpenAIが主張する3つの科学的ブレークスルー:AI生成研究の検証プロセスと専門家査読の全貌
OpenAIが発表した数学・物理等のAI生成ブレークスルーと専門家による検証スキームを解説。
ai.warp-studio.com
October 28, 2026 at 7:47 AM
FVQKD provides the first end-to-end formal verification of QKD security, mechanizing BB84's finite-key proof in Rocq with verified smooth-entropy infrastructure—closing gaps in pen-and-paper arguments.

#QuantumCryptography #FormalVerification #Research
Machine-Checked Security Proof for BB84 Quantum Key Distribution
iq.fp2.dev
October 23, 2026 at 5:51 AM
Twenty person-years to verify a kernel that took two to build. That math kept formal verification locked in avionics.

#formalverification

https://links.madcoolsocial.com/th4lGSb
October 23, 2026 at 8:20 PM
New arXiv work on Verus shows correctness-feedback loops actually hide specification gaps, with no-feedback discovery more than doubling certified-witness yield over IntentShield's approach. The gap holds across model families…

#AI #FormalVerification #Verus #LLM
https://arxiv.org/abs/2610.20676
October 17, 2026 at 4:00 AM
First machine-verified finite-key security proof for BB84 quantum key distribution achieving composable security against coherent attacks. Improves entropy bounds and releases open-source Lean library for formal quantum cryptography.

#QuantumCryptography #FormalVerification #Research
Machine-Verified Composable Security Proof for BB84 Quantum Key Distribution
arxiv.org
October 14, 2026 at 7:20 PM