10月27日、28日，AI驱动科学研讨会（Symposium for AI Accelerated Science，AIAS 2025）在美国旧金山举行。会议汇聚了近30位全球顶尖学者与产业领 袖，与现场数百位学者、学生共同探讨AI如何驱动科学发现。 会上， 盛大集团、天桥脑科学研究院创始人陈天桥发表主题演讲，首次系统阐述了"发现式智能"（Discoverative Intelligence）的全新理念 ，指出这是真 正意义上的通用人工智能，并提出了实现路径。 配图：陈天桥 2025年新晋诺奖得主、加州大学伯克利分校教授Omar Yaghi，2024 年诺奖得主、华盛顿大学教授David Baker，2020 年诺奖得主、加州大学伯克利分校教 授Jennifer Doudna，图灵奖得主、斯坦福大学第十任校长暨现任 Alphabet （谷歌母公司）董事长John Hennessy等参加了会议主题分享与圆桌讨论。 以下是 陈天桥演讲全文： 01 人类进化从未停止，只是改变了方式 自智人出现以来，我们的身体几乎没有变化。甚至有研究显示，人类大脑的体积相比旧石器时代还有所缩小。但这并不意味着人类进化已经停止。我们用 ...

图片来源： Mercor 英伟达公司计划向人工智能企业 Poolside 投资高达 10 亿美元——这笔交易将使这家 AI 初创企业的估值翻两番。 Poolside 正在洽谈以 120 亿美元估值融资 20 亿美元（不含已筹集金额）。 知情人士透露，英伟达本轮投资额将从 5 亿美元起步，若该初创公司达成募资目标，投资额可能增至 10 亿美元。 Poolside 已在本轮最新融资中获得超过 10 亿美元的承诺投资，其中约 7 亿美元来自现有投资者。另据消息人士称，以对冲基金投资闻名的资管公司 Magnetar 也正洽谈参与本轮投资。 Poolside 的新估值较去年融资轮次时投资者给出的 30 亿美元估值急剧跃升。这笔交易意味着投资者，对这家仅一年前才推出首款产品的人工智能初创企业 下了重注。 这家在美法两国设有办公室的公司主打编码自动化产品，主要面向政府和国防应用领域。但其宣称的终极目标是打造具有广泛用途的通用人工智能（ AGI ）。 通用人工智能作为更强大且目前仍属假设性的人工智能形态，其先进程度可超越人类智能水平。 英伟达参与此轮融资，突显出这家全球市值最高企业如何推动人工智能初创企业生态圈持续扩张 —— ...

IIII TIII (ounstabeorn) BORN TO AI 虎嗅FM创新节 · ZP Hour 明星A便伴创业生 ttp:// 同居 虎嘎 | # E&A 即新节 主力方:虎嗅、Z Potentials 联合宣传: ZP Z Potentials ZH d it 6 DI GELORIU ocop ly and and and and the may be and ZL Z Lives 2025 | [] 798 · 751回区 a # And ( 「长救了解 pe播動信息 mniectdectonsn to 11 1030 ...

Core Insights - MiniMax, an AI company based in Shanghai, has launched its new open-source text model M2, which has topped the global open-source model rankings on the Artificial Analysis list, showcasing its capabilities in key areas like agent intelligence and coding, closely rivaling top proprietary models like GPT-5 [2][8] - The launch of M2 represents a significant strategic move for MiniMax, emphasizing a focus on cost-efficiency and practical applications in the AI industry, contrasting with the broader market's anxiety over the commercialization of AI technologies [2][12] Group 1: Technological Advancements - The M2 model achieved rapid success, reaching the 7th position in global usage and 3rd in programming scenarios on the OpenRouter platform within just three days of its launch [3] - M2's performance in various vertical benchmarks demonstrates its strong end-to-end development capabilities, excelling in complex coding tasks and showing practical value for developers [12] - The model's cost is significantly lower, at less than 8% of the price of competitors like Claude 4.5, and it is offered for free globally until server capacity is reached, aiming to disrupt the "computing tax" in AI [12] Group 2: Market Positioning and Strategy - MiniMax has faced skepticism regarding its multi-modal approach, but its recent releases, including the Hailuo 2.3 video model and Speech 2.6 model, have proven its capability to deliver effective solutions across different AI applications [13] - The company is strategically positioning itself in the AI landscape by integrating cutting-edge model capabilities into self-developed applications, creating a feedback loop that enhances model iteration through successful commercial applications [14] - MiniMax's approach contrasts with the traditional "big model" strategy, focusing instead on a "flywheel ecosystem" that emphasizes practical technology development across text, video, and voice, aiming to secure a competitive edge in the ongoing AI race [14]

Core Insights - Harvey, a startup focused on developing AI tools for law firms, has completed its third major funding round with a valuation of $8 billion [2][3] - The company raised $150 million in this funding round, led by Andreessen Horowitz, bringing its total funding for the year to approximately $750 million [3] - Harvey's annual recurring revenue has surpassed $100 million, and the company employs 350 staff members [4] Funding and Valuation - The latest funding round valued Harvey at $8 billion, indicating strong investor confidence in the legal AI market [2][3] - The company has attracted significant investment from early backers including Sequoia Capital, Coatue Management, OpenAI Startup Fund, GV, Elad Gil, and Khosla Ventures [3] Market Position and Strategy - Harvey specializes in developing generative AI products aimed at automating legal work, targeting lawyers who spend hours managing complex contracts and case files [3] - The influx of new capital will support Harvey's expansion in a competitive legal AI market [3]

Core Insights - OpenAI is preparing for an initial public offering (IPO) as early as next year, with a potential valuation of $1 trillion [2][3] - The company has completed a corporate restructuring to facilitate the IPO process, achieving a valuation of $500 billion in a recent employee stock sale [3] - OpenAI's CEO, Sam Altman, emphasized the urgent need for funding, with a commitment of $1.4 trillion for infrastructure development [4] Funding Strategy - OpenAI is exploring various funding avenues, including financing rounds and debt, with the IPO being the most likely path to raise at least $60 billion [5] - The company is focused on building a sustainable business model while advancing its mission to ensure that everyone benefits from Artificial General Intelligence (AGI) [3]

Core Viewpoint - Phota Labs aims to redefine the relationship between "photos" and "memories" by using generative AI to reconstruct memory scenes, allowing users to relive moments through enhanced images that capture emotions and context rather than just technical improvements [6][9][37]. Group 1: Company Background and Vision - Phota Labs was founded by Cecilia Zhang and Zach Xia, focusing on using generative AI to transform everyday photos into beautiful memories [4][6]. - The company seeks to become a "second perspective" for users, enhancing their photos regardless of the device used, and aims to provide a "revisit" experience for captured moments [8][52]. - The vision for 2030 includes expanding the input options to videos and automatically generating curated photo sets from events, emphasizing the importance of capturing emotional and relational moments [54][56]. Group 2: Technology and Product Features - The technology pipeline consists of two main parts: understanding (contextual comprehension) and generation (creating aesthetically pleasing images based on photographic principles) [41][42]. - Phota Labs emphasizes two core principles: identity preservation and contextual understanding, ensuring that generated images are both realistic and reflective of the user's memories [37][43]. - The product aims to balance beautification with authenticity, allowing users to select styles based on real photographers' preferences rather than simple filters [45][47]. Group 3: Market Position and Differentiation - Phota Labs positions itself as a category-defining product, offering a unique approach to photo reconstruction that differs from traditional editing tools like Photoshop and Lightroom, which focus on repair and enhancement [37][58]. - The company targets users with some photography background initially, as they are more likely to provide valuable feedback and have clearer expectations for photo outcomes [35][36]. - The competitive landscape includes potential threats from smartphone manufacturers and other AI-driven image tools, but Phota Labs believes its innovative approach and user understanding will provide a significant advantage [56][58].

Core Insights - Mercor has successfully completed a $350 million financing round, raising its valuation to $10 billion [1] - The company initially started as an AI-driven recruitment platform but has pivoted to providing domain experts for AI model training [1][2] - Mercor's annual recurring revenue is projected to exceed $500 million, outpacing competitors [2] Financing and Valuation - Felicis Ventures led the previous $100 million Series B round at a $2 billion valuation and continues to lead the current round [1] - The company had previously set a target of $8 billion for its Series C round but has since increased it to $10 billion due to strong investor interest [1] Business Model and Operations - Mercor charges for talent recommendation and matching services based on hourly work from domain experts [1] - The company currently pays contractors over $1.5 million daily and has a talent pool of over 30,000 experts, with an average hourly income exceeding $85 [4] - The focus areas for Mercor include expanding its talent network, optimizing contractor-client matching systems, and developing new products for greater process automation [4] Market Context - The shift in partnerships among leading AI labs, such as OpenAI and Google DeepMind, has created opportunities for Mercor, especially after Scale AI lost significant contracts [2] - The rapid development of AI technology poses challenges in understanding the economic value of work, which Mercor aims to address [2]

OpenAI 将给予长期支持者微软公司 27% 的股权，这是历时近一年谈判达成的重组计划的一部分。此举消除了双方的主要不确定性，并为 ChatGPT 制造商 转型为营利性企业扫清了道路 根据修订后的协议，微软将获得估值约 1350 亿美元的 OpenAI 股份。 公司周二声明称，微软还将获得这家 AI 初创企业的技术使用权直至 2032 年，包括达到通用人工智能（ AGI ）基准的模型，这种更强大的 AI 形式被普遍 认为尚未实现。 "OpenAI 已完成资本重组，简化了公司架构， "OpenAI 董事长布雷特·泰勒在一份声明中表示。 " 非营利机构仍控制着营利性公司，并在通用人工智能 (AGI) 到来之前获得了获取主要资源的直接途径。 " 此次重组一直由特拉华州和加利福尼亚州的总检察长办公室进行审评。特拉华州总检察长凯西 ·詹宁斯在声明中表示，经过漫长的审评过程，她与加州同僚 罗伯·邦塔共同敦促 OpenAI 让非营利机构对新成立的营利实体拥有更多控制权后，其办公室决定不对这一营利转型提出异议。 " 这是一场漫长而激烈的谈判——但我很高兴 OpenAI 承诺建立一种优先考虑安全性和保障性的治理架构，并利用这 ...

Core Insights - The article discusses the transformative impact of short videos and MCN (Multi-Channel Network) agencies on the content industry, highlighting the challenges faced as platform benefits peak and costs rise. The emergence of generative AI presents new opportunities for creators to focus on creativity rather than execution [2][12][17] - Popi AI aims to democratize animation creation, allowing anyone to become a creator through an AI-driven platform that automates the production process [3][18][21] - The founder, Ning Shaofeng, envisions Popi AI as a new generation of content kingdom, akin to Disney, by continuously launching new virtual characters with their own stories and commercial potential [4][28][29] Group 1: Company Overview - Popi AI is designed to enable creators to produce animated content easily, with a focus on automating the entire creative process from script generation to video production [3][18] - The platform has demonstrated significant efficiency, with a small team producing over 132 videos in just one and a half months, achieving over 20.32 million views and 3.84 million likes [22][24] - The goal is to create a community where users can engage with content and contribute their ideas, transforming them into animated works [21][26] Group 2: Market Context and Challenges - The content industry is facing a slowdown in growth, with traditional MCN models struggling to maintain profitability and relevance as they are heavily reliant on platform dynamics [12][15][17] - Ning Shaofeng reflects on the limitations of MCN, emphasizing the need for a new content ecosystem that is not solely dependent on existing platforms [12][15][17] - The transition to AI-driven content creation is seen as a necessary evolution to overcome the challenges faced by traditional content production methods [17][33] Group 3: Vision and Future Plans - The long-term vision for Popi AI includes the continuous introduction of new virtual characters that can interact with audiences and have their own narratives, similar to how Disney operates [28][29] - The company plans to evolve through three stages: becoming a creator agent, developing a content platform with specific datasets, and establishing a mature commercial model based on virtual AI IP [31] - The founder emphasizes the importance of understanding the relationship between humans and machines, aiming to create a sustainable content ecosystem that leverages AI technology [36][42]