以下文章来源于智东西 ，作者王 涵 智东西 . 智能产业新媒体！智东西专注报道人工智能主导的前沿技术发展，和技术应用带来的千行百业产业升 级。聚焦智能变革，服务产业升级。 将投中网设为"星标⭐"，第一时间收获最新推送 安德鲁·塔洛克，现任Thinking Machines Lab联合创始人兼首席研究员。 OpenAI前CTO带队"团拒"。 作者丨 王涵 编辑丨 漠影 来源丨 智东西 智东西8月4日消息，据外媒《华尔街日报》近日报道，知情人士透露，与OpenAI前首席技术官米拉 ·穆拉蒂（Mira Murati）联合创办Thinking Machines Lab的安德鲁·塔洛克（Andrew Tulloch），拒绝了扎克伯格的可能高达 15亿美元（约合人民币108.2亿元） 的薪酬。 据报道，试图收购Thinking Machines Lab失败后，扎克伯格在随后数周内接触了该公司 约十几名 员工，最终成功 带走0个人 。 《华尔街日报》报道称，Meta发言人安迪·斯通（Andy Stone）回应称该薪酬描述"失实且荒谬"， 强调任何薪酬方案都需以股价上涨为前提，并声明Meta无意收购Thinking Mach ...

在下一代大模型GPT-5备受期待之际，一项名为"通用验证器"的新技术正浮出水面，揭示了OpenAI可能 用于拉开竞争差距的"秘密武器"。 OpenAI的"通用验证器"或将直接影响GPT-5模型的市场竞争力，8月4日据科技媒体The Information援引 知情人士消息报道，这项技术已被应用于GPT-5的开发过程中。 该技术的核心机制，被比作一场"证明者-验证者游戏"。简而言之，它让一个AI模型扮演"验证者"的角 色，去检查和评判另一个"证明者"模型生成的答案。通过这种内部对抗和反馈，系统性地提升模型的输 出质量。这一自动化流程旨在解决强化学习（RL）在创意写作等主观领域或数学证明等复杂领域难以 验证的瓶颈。 OpenAI内部研究人员已在社交平台X上间接证实了相关方法的有效性。研究员Noam Brown表示，这些 技术是"通用的"，能让大模型"在难以验证的任务上表现得更好"。这也标志着OpenAI正试图攻克AI商 业化应用中的核心痛点——可信度。 "证明者-验证者"的对抗游戏 "通用验证器"的技术细节，最早在OpenAI于2024年7月发表的一篇题为《证明者-验证者游戏提升大语言 模型可读性》的论文中被阐述 ...

AI产业发展及应用概况—产业政策 国家大力支持算力、算法、数据、应用等AI核心环节发展，产业政策体系持续完善 生成 式AI 基础 设施 未来 产业 标准 化 安全 治理 AI 教育 国家网信办等七部门 2023/07 《生成式人工智能服务管理暂行办法》 工信部等六部门 2023/10 《算力基础设施高质量发展行动计划》 工信部、中国科学院等七部门 2024/01 《关于推动未来产业创新发展的实施意见》 工信部、国家发改委等六部门 2024/06 《国家人工智能产业综合标准化体系建设指南 （2024版）》 全国网络安全标准化技术委员会 2024/09 《人工智能安全治理框架》1.0版 气象 应用 内容 标识 数据 标注 专利 申请 2025/04 中国气象局、国家网信办. 《人工智能气象应用服务办法》 2025/03 国家网信办、工信部等四部门 《人工智能生成合成内容标识办法》 2024/12 国家发改委、国家数据局等四部门 《关于促进数据标注产业高质量发展的实施意见》 2024/11 国家知识产权局 《人工智能相关发明专利申请指引（试行）》 2024/11 教育部办公厅 《关于加强中小学人工智能教育的通知》 ...

当地时间8月4日，多家非政府组织组建联盟发起公开信，要求OpenAI就计划中的公司重组提高透明 度。这封由Encode、The Midas Project、EyesOnOpenAI及加州非营利、慈善和劳工组织联盟等共同推 动的公开信，已获包括四位诺贝尔奖得主在内的众多知名人士及组织签署，信中集中表达了对OpenAI 可能弱化非营利使命、优先追逐利润的深切担忧。 签署者包括四位诺贝尔奖得主，包括有"AI教父"之称的多伦多大学名誉教授杰弗里・辛顿（Geoffrey Hinton，2024年物理学奖得主），罗马萨皮恩扎大学名誉教授乔治・帕里西（Giorgio Parisi，2021年物 理学奖得主），哈佛大学教授奥利弗・哈特（Oliver Hart，2016年经济学奖得主），以及哈佛大学及波 士顿大学名誉教授谢尔登・李・格拉肖（Sheldon Lee Glashow，1979年物理学奖得主）。 其他知名人士中，包括以太坊联合创始人维塔利克・布特林（Vitalik Buterin），牛津大学人工智能伦理 研究所研究员唐凤（Audrey Tang），哈佛大学法学院教授劳伦斯・莱斯格（Lawrence Lessig），加 ...

Group 1: China-Europe Railway Express - The transit time for the China-Europe Railway Express via the "Middle Corridor" to Turkey has been reduced to 15 days [5][6][7] - The route starts from China, passing through Kazakhstan, the Caspian Sea, Azerbaijan, Georgia, and extending to Turkey and European countries [8] Group 2: US Tariff Measures - The US has announced a 50% tariff on imported copper semi-finished products and high-copper derivatives starting August 1 [9][10] - From August 29, the US will suspend the tax exemption for imported packages valued at $800 or less [12] - A 40% tariff will be imposed on products imported from Brazil starting August 6, raising the total tariff rate for most Brazilian products to 50% [12] Group 3: South Africa's Economic Impact - The South African government warns that the US's 30% tariff on South African goods could lead to a loss of approximately 30,000 jobs [13][14] - The high tariffs are expected to severely impact South Africa's automotive manufacturing and agricultural processing sectors [14] Group 4: EU AI Regulation - The EU AI Act has come into effect, imposing fines of up to €35 million for non-compliance, with general AI systems also under regulation [17][18] - Member states are required to appoint market regulatory bodies to oversee compliance with the AI Act [17] Group 5: OpenAI Funding - OpenAI has raised over $8 billion in its latest funding round, achieving a valuation of $300 billion [20][21] - The funding round was led by Dragoneer Investment Group, with participation from several major investment firms [21] Group 6: Meta's Data Center Asset Sale - Meta plans to sell $2 billion worth of data center assets to share the costs of AI infrastructure [25][26] - The company is exploring partnerships with financial entities to co-develop data centers [26] Group 7: Foxconn's Shift to AI - Foxconn is selling its electric vehicle factory in Ohio for approximately $375 million to focus on AI data centers [36][37] - The company aims to pivot its business strategy towards AI infrastructure amid a downturn in the electric vehicle market [39] Group 8: South Korea's AI Semiconductor Initiative - The South Korean government has launched a project worth 30 billion KRW to support the optimization of AI semiconductor design [43][44] - The initiative aims to enhance domestic AI semiconductor companies' participation in national AI projects [43] Group 9: Quantum Research Awards - The 2025 "Mozi Quantum Award" has been awarded to three researchers in the field of quantum simulation for their groundbreaking work [46] Group 10: TSMC's 2nm Technology Transfer - TSMC is preparing to transfer its 2nm technology to the US, with a new production line being set up in Arizona [51][52] - This move is part of a broader trend among tech companies to establish manufacturing capabilities in the US [53]

明敏 发自 凹非寺 量子位 | 公众号 QbitAI GPT-5又咕咕，但是把Claude新模型诈了出来—— Claude Opus 4.1 ，被曝正在进行内部测试。 文件将这个模型描述为"更具问题解决能力"，推测它可能会 重点提高推理和规划能力 。 有网友补充说，最近使用Claude Code时会被询问使用体验，可能是在进行一些A/B测试。 尽管从型号名称判断，4.1会是一次小版本更新，但参考此前从Claude-3到Claude-3.5的"飞升"，以及同样惊艳的Claude-3.7……历史重演的 话，0.1版本升级也可能是重大飞跃。 网友们似乎也不担心Claude-4.1的性能不够好，大家更在意的是模型价格太高怎么办。 毕竟，Claude模型的生产力是有目共睹的， 尤其是在编程方面 。 最新数据显示，Anthropic过去7个月的ARR （年度经常性收入） 翻了5倍，涨到50亿美元。 其中API收入主要来自于编程，两个最大客户Cursor和GitHub Copilot，就带来了14亿美元的收入。 Claued-4.1箭在弦上？ 被曝光的模型型号为"claude-leopard-v2-02-prod"，它在配 ...

Core Viewpoint - The upcoming AI chess competition aims to showcase the performance of various advanced AI models in a competitive setting, utilizing a new benchmark testing platform called Kaggle Game Arena [2][12]. Group 1: Competition Overview - The AI chess competition will take place from August 5 to 7, featuring eight cutting-edge AI models [2][3]. - The participating models include notable names such as OpenAI's o4-mini, Google's Gemini 2.5 Pro, and Anthropic's Claude Opus 4 [7]. - The event is organized by Google and aims to provide a transparent and rigorous testing environment for AI models [6][8]. Group 2: Competition Format - The competition will follow a single-elimination format, with each match consisting of four games. The first model to score two points advances [14]. - If a match ends in a tie (2-2), a tiebreaker game will be played, where the white side must win to progress [14]. - Models are restricted from using external tools like Stockfish and must generate legal moves independently [17]. Group 3: Evaluation and Transparency - The competition will ensure transparency by open-sourcing the game execution framework and environment [8]. - The performance of each model will be displayed on the Kaggle Benchmarks leaderboard, allowing real-time tracking of results [12][13]. - The event is designed to address the limitations of current AI benchmark tests, which struggle to keep pace with the rapid development of modern models [12].

2025上半年AI核心成果及趋势报告 2025.7 量子位智库 QbitAI Insights 分析师 Xuanhao xuanhao@qbitai.com 目录 01 应用趋势 序⾔ 2 • ⼈⼯智能可能是⼈类有史以来最重要的发明，我们也正在见证AI的飞速发展，技术突破与场景落地交织推动着⾏业加速演进。 量⼦位智库将在本报告中为⼤家介绍2025年上半年，AI领域的关键动态和趋势，旨在为决策者、从业者和创新者提供前沿洞 察，帮助他们在竞争激烈的⽣态中把握机遇。本报告将从应用、模型、技术、⾏业4个⽅面进⾏展开： • 应用趋势：包括通用类Agent开始进⼊主流、垂类Agent开始涌现、AI编程获得市场验证⾼速增长、模型上下⽂协议（MCP）获 得⾏业关注等应用侧核⼼变化 • 模型趋势：包括推理模型能⼒进步、⼯具使用能⼒落地、模型多模态能⼒增强、小模型加速应用普及、模型评估加速演化等 模型层面的核⼼变化 • 技术趋势：包括模型不同训练阶段的重⼼变化、强化学习的重要性、多智能体（Multi-Agent）系统和在线学习的优势、新型 模型架构迭代和应用情况等技术范式的核⼼变化 • ⾏业趋势：主要介绍AI领域的⾏业核⼼动态， ...

Summary of Key Points from the Conference Call Industry: Artificial Intelligence (AI) Core Insights and Arguments - **Introduction of GPT-5**: OpenAI's upcoming model, GPT-5, is generating positive early feedback, particularly in coding tasks, which is a critical area for the company [3][4][5] - **Performance Improvements**: GPT-5 shows enhanced capabilities in various domains, especially in software engineering, outperforming previous models and rival Anthropic's Claude Sonnet 4 in specific tests [7][10] - **Integration of Models**: The model aims to combine traditional large language models (LLMs) with reasoning models, allowing users to control the reasoning capabilities based on task complexity [5][6] - **Practical Applications**: GPT-5 is better equipped to handle real-world programming challenges, such as modifying complex legacy code, which has been a historical weakness for OpenAI's models [8][9] - **Market Implications**: The success of GPT-5 could significantly impact OpenAI's business and its competitors, as coding assistants powered by Anthropic's models are projected to generate substantial revenue for Anthropic [10][12] Additional Important Content - **Caveats on Model Understanding**: There is uncertainty regarding the exact nature of GPT-5, with speculation that it may function as a router directing queries rather than a single, unified model [13] - **Future Improvements**: Experts suggest that future advancements may stem more from post-training reinforcement learning rather than scaling up pretraining processes [15][17] - **Investor Sentiment**: OpenAI executives are optimistic about the potential for future models, claiming they can reach "GPT-8" using current model structures [17] Implications for Stakeholders - **Impact on Suppliers and Investors**: Strong performance of GPT-5 is seen as beneficial for OpenAI's chip supplier Nvidia and data center firms, as well as for equity and debt investors concerned about AI development trajectories [12]

Ask OpenAI CEO Sam Altman. Art by Mike Sullivan Exclusive Inside OpenAI's Rocky Path to GPT-5 The troubles OpenAI has faced in developing GPT-5 point to slowing AI progress across the industry. Researchers believe advances in reinforcement learning will help to overcome that. By Stephanie Palazzolo, Erin Woo and Amir Efrati OpenAI made waves across the industry in December when it published the results from its tests of artificial intelligence that performs better on tasks when it gets more time and computi ...