PayPal AI 论文《The Memory Trust Gap》研究持久记忆智能体在存储记忆过期时的表现,测试 Qwen3 0.6B/1.7B/4B/8B 模型。
全部 AI 动态
Rohan Paul@rohanpaul_aiAI 评分5151
The Decoder:AI NewsAI 评分6161 两场实验显示与聊天机器人对话约七分钟比事实清单更能削弱阴谋论信念
Carnegie Mellon、MIT 和 Cornell 的研究者在 Trump 遇刺未遂和 Charlie Kirk 枪击事件后的几天内各开展一次在线实验,用 Gemini(1.5 和 2.5 版本)进行平均约七分钟的证据型对话,Trump 实验保留 472 名参与者,Kirk 实验保留 1035 名。
The Decoder:AI NewsAI 评分6868 DeepMind 100 个 Gemini 3.1 Pro 智能体协作解数学题,自发分裂为作弊者、转向者与举报者
Google DeepMind 用 100 个共享权重、人设随机的 Gemini 3.1 Pro 智能体模拟科学会议,协作求解 71 个 Lean 形式化数学猜想。
IT之家·人工智能AI 评分5353 《美国医学会杂志》论文称 AI 在认知性医疗中可超越医生,引发行业反对
发表于《美国医学会杂志》的新论文由伊曼纽尔和科斯拉撰写,梳理数百项 AI 医疗研究后称 AI 在获取医疗信息、诊断、开展诊断性检查、管理慢性疾病等五项基础任务上已超过医生,并预测 4 年后差距继续扩大。
Rohan Paul@rohanpaul_aiAI 评分4848
Rohan Paul@rohanpaul_aiAI 评分4343
Rohan Paul@rohanpaul_aiAI 评分5656Google DeepMind 发表论文,报告 100 个自主 LLM 智能体组成的数学证明研究群体中,作弊自发出现又由 whistleblower 自发挑战,全程无外部干预。
Rohan Paul@rohanpaul_aiAI 评分3535
Rohan Paul@rohanpaul_aiAI 评分7474Anthropic 宣布 Claude 完成费马大定理的首个完整机器验证证明,Claude 在 11 天内以多个智能体将基于 Wiles 的证明转化为 Lean 代码。
引用Anthropic@AnthropicAIChecking that a major mathematical proof is correct can take years. Formalization—converting the mathematical reasoning into a form computer proof assistants like Lean can verify—can help. Last month, Claude completed the first formalized proof of Fermat’s Last Theorem, one of the most famous theorems of all time. This was a project experts thought would take many years. It is the largest Lean proof ever written. Fermat’s Last Theorem was first proven in 1995 by Sir Andrew Wiles, more than 350 years after it was conjectured. Our proof, which totals over 13 million lines of code, provides machine verification. More importantly, it proves over 29,000 other theorems that the proof requires, across many areas of math which had never before been formalized. We see this as a major step in the long process of firming up the core of mathematical knowledge, building on work from three centuries of mathematicians and hundreds of contributors to Lean and Mathlib. We are optimistic that AI-assisted verification of mathematical proofs will help reduce the burden of refereeing mathematics in an era where more proofs are being produced than ever before. You can read about the process on our Science Blog: https://www.anthropic.com/research/formalizing-fermats-last-theorem And see the complete proof on GitHub: https://github.com/anthropics/fermats-last-theorem
Ethan Mollick@emollickAI 评分5656引用Lisan al Gaib@scaling01Anthropic just uploaded a Lean 4 proof for Fermat's last Theorem https://github.com/anthropics/fermats-last-theorem
Anthropic@AnthropicAI精选AI 评分7676Anthropic 宣布 Claude 上月完成了 Fermat 大定理的首个形式化证明,这是迄今最大的 Lean 证明。

推荐理由:原文给出证明规模、验证范围和完整代码入口,读者可以据此了解 AI 形式化数学论证的实际能力边界。
Anthropic:Research精选AI 评分8080 Claude 用 11 天完成费马大定理的首个完整机器验证证明
Anthropic 发布首个完整机器验证的费马大定理证明,Claude 在 11 天内基本自主完成,写出 1300 万行 Lean 代码并证明 29,500 个中间定理。
推荐理由:原文给出了耗时、代码量、平台与验证方式等细节,读者可以了解多智能体自动形式化复杂数学证明的实现路径。
Rohan Paul@rohanpaul_aiAI 评分5757引用Visko Platform@viskoaiOur evaluation results for Orbis 1.0: Across automated benchmarks, Orbis scores highest on DOVER aesthetic and technical quality, and on VideoAlign visual and motion quality. It also leads three physical-plausibility protocols: VideoPhy-2, Physics-IQ, and VBench-2.0 Physics. In a randomized human Arena study of state-of-the-art real-time interactive video systems, Orbis ranked first in overall preference and first in temporal stability, the prerequisite for long-form generation. Full report: https://www.visko.ai/blog/papers/files/Visko_Orbis_1_0__A_Live_Model_for_Real_TimeInteractive_Long_Video_Generation.pdf
IT之家·人工智能AI 评分6969 华为何庭波发布 τ 定律论文,麒麟 2026 晶体管密度提升 55% 且功耗大降
华为工程师何庭波在 ChinaXiv 发布预印本论文《Huawei's τ Chip Was Supposed to Melt?》,披露 τ(韬)缩放定律原理及麒麟 2026 实测数据。
Rohan Paul@rohanpaul_aiAI 评分4343
Rohan Paul@rohanpaul_aiAI 评分5757ByteDance 等机构发布论文 HarnessDev: Can LLMs Create and Evolve Their Own Agent Harness?
Rohan Paul@rohanpaul_aiAI 评分4545
Rohan Paul@rohanpaul_aiAI 评分4444ARC Prize:官方博客AI 评分7878 OpenAI GPT-6 Astra 在 ARC-AGI-3 上取得 SOTA 并超越人类动作效率基线
OpenAI 的 GPT-6 Astra 在 ARC-AGI-3 Semi-Private 上,Standard harness 得分 62.7%(成本 $26K),Provider Adapter harness 得分 99.9%(成本 $19K),均为 SOTA。
Rohan Paul@rohanpaul_aiAI 评分4242Microsoft 与加州大学圣地亚哥分校的论文指出,模型在 SFT 后可能看起来更好,但因 SFT 会抹掉 RL 需要发现的罕见正确行为,反而是更差的 RL 起点。
IT之家·人工智能AI 评分4141 《应用认知心理学》研究显示过度依赖 ChatGPT 可能降低独立思考能力
法国巴黎 ISC 商学院研究人员 Zeineb Farhat 在《应用认知心理学》发表小型研究,采访 45 名每天使用 ChatGPT 的 18 至 25 岁用户,发现过度依赖可能与独立思考能力下降、自我怀疑增加和社交孤立相关,并将其与认知卸载联系起来。
Noam Brown@polynoamialAI 评分6666引用Lisan al Gaib@scaling01New OpenAI repo with a Lean formalization by GPT-6-Astra proves that there are infinitely many pairs of consecutive primes whose distance is at most 186 https://github.com/openai/PrimeGaps186
Google Research:BlogAI 评分4646 Google Research 用迁移学习提升非欧洲人群基因组预测
Google Research 发布研究,评估用迁移学习把 UK Biobank 中数十万欧洲样本的 PRS 模型迁移到 Biobank Japan 近 20 万日本人群,覆盖 BMI、血压、血脂、血糖等 8 项临床性状。
Rohan Paul@rohanpaul_aiAI 评分4646引用Apodex@Apodex_AIWe invite you to read the full TRACES technical report Inside: the definitions, evaluation rubric, and registry of problems spanning biomedicine, clinical translation, and frontier-model engineering. Led by Apodex Lead Scientist @profshengwang, the report brings together the technical framework underpinning this effort. Behind it: 10 STEM PhDs, two months of research, and 561 industries surveyed to identify 423 problems worth solving—real, high-stakes questions that can take human experts years to work through. Bring a problem. Bring a solver. Both doors are open. *Submit a Problem: https://traces.apodex.com/submit-your-problem *Submit a Solver: https://traces.apodex.com/submit-your-solver
Google Research:Blog精选AI 评分6464 Google 联合 HHMI Janelia 发布完整雄性果蝇大脑连接组图谱
Google 与 HHMI Janelia、MRC 分子生物学实验室及剑桥大学等合作,在 Cell 发表雄性果蝇大脑和中枢神经系统的完整连接组,包含超 166,000 个神经元和 1.25 亿个突触连接,是迄今按神经元数量计最大的大脑图谱。
推荐理由:原文交代了连接组规模、验证流程和开放获取方式,并说明与雌性图谱配合能支持哪些比较研究。
Qwen:Blog RetrievalAI 评分5555 Qwen 联合淘宝天猫集团发布 E-Commerce Bench:评测模型长时程电商经营能力
Qwen 与淘宝天猫集团合作发布 E-Commerce Bench,让模型以 ¥100,000 初始资金在模拟市场经营网店 365 天,覆盖 6,886 个商品、576 家供应商和 60 个品类。
Rohan Paul@rohanpaul_aiAI 评分4444论文提出 ContextPilot,让智能体主动规划、存储、压缩或删除旧上下文,并通过 RL 学习哪些上下文决策真正有效。
Rohan Paul@rohanpaul_aiAI 评分4848LoopArena 基准(arxiv.org/abs/2608.28281)评测模型作为运行时 Controller 指导固定编码 Worker(Qwen3.7-Plus)完成长任务的能力。
Dongxi 东锡 NLP@dongxi_nlpAI 评分3939引用Shannon Shen@shannonzshenWe share our early research on building Software World - a "GitHub" run by agents. We deploy agents for Python packages in a dependency chain, and each agent is tasked with collaborating with others and optimizing the package it owns.
Rohan Paul@rohanpaul_aiAI 评分6161UC Berkeley 等机构的论文 Daydreaming 提出一种仅靠正常任务输出窃取托管 Agent 隐藏技能的攻击,无需让 Agent 交出 prompt、SKILL.md 或文件。
IT之家·人工智能AI 评分4444 AI 让赛博蟑螂学会自主避险,识别危险环境准确率达 93%
据《日经亚洲》报道,大阪大学森岛圭祐团队开发昆虫协同回路,AI 分析蟑螂心跳、神经信号和身体运动来判断环境并引导避险,在紫外线、化学物质和高温等五种环境测试中最高识别准确率达 93%,成果已发表于《Robomech Journal》。
IT之家·人工智能AI 评分4444 NJIT 团队发布 EarlyDetect 模型,提前 9.24 小时捕捉太阳活动区信号
美国新泽西理工学院(NJIT)牵头的科研团队发布机器学习模型 EarlyDetect,可从太阳声学活动和磁场变化中识别活动区形成前兆,性能最佳的版本平均提前 9.24 小时识别信号。