Humans, AIs, and other things evolved via optimization and selection pressures can't actually understand things, they can only mimic an understanding of things. True understanding is accessible only to symbolic formal systems like lean4, clojure, and PHP.
November 5, 2026 at 12:53 AM
Humans, AIs, and other things evolved via optimization and selection pressures can't actually understand things, they can only mimic an understanding of things. True understanding is accessible only to symbolic formal systems like lean4, clojure, and PHP.
#AI #agents could prove answers
Explore Lean4: the intersection of #neuro-symbolic reasoning, #higher-order #logic, & #knowledgegraphs, combining AI-driven discovery with formal #verification to make #agent #reasoning #trustworthy and accountable.
From claims to evidence. From answers to proofs.
Explore Lean4: the intersection of #neuro-symbolic reasoning, #higher-order #logic, & #knowledgegraphs, combining AI-driven discovery with formal #verification to make #agent #reasoning #trustworthy and accountable.
From claims to evidence. From answers to proofs.
Lean 4 Knowledge — Graphs, logic & certified reasoning
Lean 4 for graph engineering, typed ontologies, and certified reasoning. Explore projects, articles, mini-books, and courses by Volodymyr Pavlyshyn.
lean4knowledge.com
November 2, 2026 at 2:44 PM
#AI #agents could prove answers
Explore Lean4: the intersection of #neuro-symbolic reasoning, #higher-order #logic, & #knowledgegraphs, combining AI-driven discovery with formal #verification to make #agent #reasoning #trustworthy and accountable.
From claims to evidence. From answers to proofs.
Explore Lean4: the intersection of #neuro-symbolic reasoning, #higher-order #logic, & #knowledgegraphs, combining AI-driven discovery with formal #verification to make #agent #reasoning #trustworthy and accountable.
From claims to evidence. From answers to proofs.
she has to give a talk in lean4
October 28, 2026 at 8:29 AM
she has to give a talk in lean4
今「LEAN4の脆弱性」を発見したら、
世界転覆出来るだろ。
世界転覆出来るだろ。
October 24, 2026 at 3:55 AM
今「LEAN4の脆弱性」を発見したら、
世界転覆出来るだろ。
世界転覆出来るだろ。
Formal verification meets agentic RL: new framework grounds agent training in Lean4 proofs, achieving 77% on terminal tasks and beating comparable models by 5-10%.
https://arxiv.org/abs/2610.22448
#AI #MachineLearning
https://arxiv.org/abs/2610.22448
#AI #MachineLearning
October 23, 2026 at 11:30 AM
Formal verification meets agentic RL: new framework grounds agent training in Lean4 proofs, achieving 77% on terminal tasks and beating comparable models by 5-10%.
https://arxiv.org/abs/2610.22448
#AI #MachineLearning
https://arxiv.org/abs/2610.22448
#AI #MachineLearning
Btw, are you using Lean4? Trying to guess the best tooling for rust. Coq+C is mature, but I would love a more modern approach P
October 23, 2026 at 7:20 AM
Btw, are you using Lean4? Trying to guess the best tooling for rust. Coq+C is mature, but I would love a more modern approach P
New Lean4AgenticRL framework uses Lean4 formalization to ground multi-level process rewards for agentic RL in software engineering tasks, addressing the limits of outcome-only rewards.
#AI #ML #DevTools
https://arxiv.org/abs/2610.22448
#AI #ML #DevTools
https://arxiv.org/abs/2610.22448
October 19, 2026 at 8:01 AM
New Lean4AgenticRL framework uses Lean4 formalization to ground multi-level process rewards for agentic RL in software engineering tasks, addressing the limits of outcome-only rewards.
#AI #ML #DevTools
https://arxiv.org/abs/2610.22448
#AI #ML #DevTools
https://arxiv.org/abs/2610.22448
【2026年版】Lean 4・形式証明の入門におすすめ！全26記事を無料Bookにまとめました | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-free-book-announcement
https://zenn.dev/dantarg/articles/lean4-free-book-announcement
【2026年版】Lean 4・形式証明の入門におすすめ！全26記事を無料Bookにまとめました
zenn.dev
October 19, 2026 at 1:48 AM
【2026年版】Lean 4・形式証明の入門におすすめ！全26記事を無料Bookにまとめました | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-free-book-announcement
https://zenn.dev/dantarg/articles/lean4-free-book-announcement
中学生でもわかる Lean 4 #25｜AI時代の「確かめ方」 | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-ai-age-verification
https://zenn.dev/dantarg/articles/lean4-ai-age-verification
中学生でもわかる Lean 4 #25｜AI時代の「確かめ方」
zenn.dev
October 17, 2026 at 5:33 PM
中学生でもわかる Lean 4 #25｜AI時代の「確かめ方」 | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-ai-age-verification
https://zenn.dev/dantarg/articles/lean4-ai-age-verification
AI math headlines everywhere, but how do we measure progress? A solved problem is useless as a benchmark.
SqueezeBench asks instead for better certified bounds, proved in #Lean4.
A success raises the bar instead of removing the problem.
github.com/leansolving/...
#AIforMath
SqueezeBench asks instead for better certified bounds, proved in #Lean4.
A success raises the bar instead of removing the problem.
github.com/leansolving/...
#AIforMath
October 16, 2026 at 5:42 PM
AI math headlines everywhere, but how do we measure progress? A solved problem is useless as a benchmark.
SqueezeBench asks instead for better certified bounds, proved in #Lean4.
A success raises the bar instead of removing the problem.
github.com/leansolving/...
#AIforMath
SqueezeBench asks instead for better certified bounds, proved in #Lean4.
A success raises the bar instead of removing the problem.
github.com/leansolving/...
#AIforMath
Не удивлюсь, если в **Toyota** пишут на **Lean4** :blobcatphotobomb:
Lean がこれだけ流行ってしまうと、Lean に何をどこまで背負わせていいか理解している人の割合は減ってゆく
October 16, 2026 at 1:58 PM
Не удивлюсь, если в **Toyota** пишут на **Lean4** :blobcatphotobomb:
IneqAgent is an agentic framework combining LLM reasoning with symbolic computation to prove mathematical inequalities in Lean4, using Monte Carlo Tree Search to coordinate lemma construction, goal reformulation, and backend…
#AI #LLMs #FormalMethods #MathAI
https://arxiv.org/abs/2610.20681
#AI #LLMs #FormalMethods #MathAI
https://arxiv.org/abs/2610.20681
October 17, 2026 at 4:01 AM
IneqAgent is an agentic framework combining LLM reasoning with symbolic computation to prove mathematical inequalities in Lean4, using Monte Carlo Tree Search to coordinate lemma construction, goal reformulation, and backend…
#AI #LLMs #FormalMethods #MathAI
https://arxiv.org/abs/2610.20681
#AI #LLMs #FormalMethods #MathAI
https://arxiv.org/abs/2610.20681
中学生でもわかる Lean 4――とっても優しい形式証明入門 | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/books/lean4-proof-literacy
https://zenn.dev/dantarg/books/lean4-proof-literacy
中学生でもわかる Lean 4――とっても優しい形式証明入門
Lean 4が初めてでも、身近な疑問と小さな例から「証明する」「確かめる」を学べる入門書。反例、前提、定義、形式証明の読み方を一歩ずつたどり、AIの答えや「形式検証済み」という説明について、何が確認され、どこから先はまだ分からないかを見極めます。証明の結果と現実の判断、決める権限
zenn.dev
October 17, 2026 at 9:46 PM
中学生でもわかる Lean 4――とっても優しい形式証明入門 | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/books/lean4-proof-literacy
https://zenn.dev/dantarg/books/lean4-proof-literacy
中学生でもわかる Lean 4 #24｜論破から検証へ | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-debate-to-verification
https://zenn.dev/dantarg/articles/lean4-debate-to-verification
中学生でもわかる Lean 4 #24｜論破から検証へ
zenn.dev
October 16, 2026 at 5:00 PM
中学生でもわかる Lean 4 #24｜論破から検証へ | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-debate-to-verification
https://zenn.dev/dantarg/articles/lean4-debate-to-verification
Proving math inequalities is hard in a specific way: you need intuition and exact computation, and they refuse to live in the same system. IneqAgent stops pretending otherwise.
https://barlark.com/ineqagent-proves-inequalities-in-lean4-and-barlark-has-thoughts/
https://barlark.com/ineqagent-proves-inequalities-in-lean4-and-barlark-has-thoughts/
October 16, 2026 at 7:00 AM
Proving math inequalities is hard in a specific way: you need intuition and exact computation, and they refuse to live in the same system. IneqAgent stops pretending otherwise.
https://barlark.com/ineqagent-proves-inequalities-in-lean4-and-barlark-has-thoughts/
https://barlark.com/ineqagent-proves-inequalities-in-lean4-and-barlark-has-thoughts/
中学生でもわかる Lean 4 #23｜人と意見を分ける | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-person-vs-claim
https://zenn.dev/dantarg/articles/lean4-person-vs-claim
中学生でもわかる Lean 4 #23｜人と意見を分ける
zenn.dev
October 15, 2026 at 12:10 PM
中学生でもわかる Lean 4 #23｜人と意見を分ける | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-person-vs-claim
https://zenn.dev/dantarg/articles/lean4-person-vs-claim
中学生でもわかる Lean 4 #22｜間違えることと負けることを分ける | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-error-not-defeat
https://zenn.dev/dantarg/articles/lean4-error-not-defeat
中学生でもわかる Lean 4 #22｜間違えることと負けることを分ける
zenn.dev
October 14, 2026 at 2:05 PM
中学生でもわかる Lean 4 #22｜間違えることと負けることを分ける | 小野 昭久(PN:古明地ゆとり/ID:dantarg) #zenn
https://zenn.dev/dantarg/articles/lean4-error-not-defeat
https://zenn.dev/dantarg/articles/lean4-error-not-defeat
Any doctor who isn't offering independent certifications in Lean4, Rocq, and Cubical Agda is a quack.
I would take the Lean-certified cancer therapy
October 11, 2026 at 10:09 PM
Any doctor who isn't offering independent certifications in Lean4, Rocq, and Cubical Agda is a quack.
October 12, 2026 at 12:50 PM
A critical look at Lean4
www.lesswrong.com/posts/zHhGa3...
www.lesswrong.com/posts/zHhGa3...
A critical look at Lean4 — LessWrong
Lots of the discussion around Lean certificates and trust centers around general philosophical ideas of how much humans should trust formal certifica…
www.lesswrong.com
October 12, 2026 at 11:34 AM
A critical look at Lean4
www.lesswrong.com/posts/zHhGa3...
www.lesswrong.com/posts/zHhGa3...