智东西2月10日消息，刚刚，马斯克旗下AI独角兽xAI的联合创始人吴宇怀在X上发文，宣布自己已从xAI离职。 ▲吴宇怀官宣推文（来源：X） 吴宇怀是一位出生于1995年的杭州学霸，初中就读于杭州建兰中学，后转至加拿大读高中。2015年他以满绩点毕业于加拿大纽布伦斯威克大 学，后在多伦多大学师从"AI教父"杰弗里·辛顿获得博士学位。 2021年博士毕业后，吴宇怀又前往斯坦福大学，在斯坦福CRFM主任Percy Liang与Jay McClelland教授的指导下完成了博士后研究。求学期间， 他还曾在谷歌DeepMind AlphaGo团队和OpenAI有过实习经历。 在进行博士后研究的同时，吴宇怀加入了谷歌，在谷歌工作至2023年。而后，他与马斯克等11位创始人，联合创办了xAI。 ▲吴宇怀工作经历（来源：领英） 吴宇怀的主要研究方向是打造具备推理能力的机器。在校期间，他主导或深度参与了多个知名项目，如自训练增强推理模型STAR、定理证明 器Alpha Geometry以及语言模型Minerva，目标是创造一个自动化的AI"数学家"。 吴宇怀在xAI的核心任务是将他之前在数学推理领域的深厚积累，应用到Grok模 ...

智东西 作者 | 王涵 编辑 | 冰倩 智东西2月10日消息，刚刚，马斯克旗下AI独角兽xAI的联合创始人吴宇怀在X上发文，宣布自己已从xAI离职。 ▲吴宇怀官宣推文（来源：X） 吴宇怀是一位出生于1995年的杭州学霸，初中就读于杭州建兰中学，后转至加拿大读高中。2015年他以满绩点毕业于加拿大纽布伦斯威克大 学，后在多伦多大学师从"AI教父"杰弗里·辛顿获得博士学位。 2021年博士毕业后，吴宇怀又前往斯坦福大学，在斯坦福CRFM主任Percy Liang与Jay McClelland教授的指导下完成了博士后研究。求学期间， 他还曾在谷歌DeepMind AlphaGo团队和OpenAI有过实习经历。 在进行博士后研究的同时，吴宇怀加入了谷歌，在谷歌工作至2023年。而后，他与马斯克等11位创始人，联合创办了xAI。 ▲从左到右分别为：Igor Babuschkin、Jimmy Ba、吴宇怀、Elon Musk（来源：YouTube） 自2024年初至今，xAI已有多位核心联创相继离职，Kyle Kosic（24年4月）、Christian Szegedy（25年2月）和Igor Babuschkin（25 ...

最近 Cowork 爆火，都说 AI 能自己干活了，那取代个搜索引擎和新闻网站应该是手拿把掐吧。（作者正在办公室瑟瑟发抖：别取代我啊！！） 那如果只用 AI 看新闻，会发生什么？ 去年九月的每个早晨，魁北克大学新闻学教授 Jean-Hugues Roy 都会做同一件事：打开七个 AI 聊天机器人，用法语问它们今天魁北克最重要的五条新闻 是什么。 一个月下来，839 条回复记录在案。 结论是：AI 在表演专业这方面真挺专业的。 原作者记录的每天早上生成式人工智能工具给出的回复的 Google 表格文件（法语）： https://docs.google.com/spreadsheets/d/1yc2Gw-6BXem45dF7jP-bsxVC7hEmN_mlaNncN1PFjmw/edit?gid=0 #gid=0 废物点心 第一次看 ChatGPT 或 Perplexity 给出的新闻摘要，你多半会觉得挺靠谱。结构清晰，段落分明，关键信息用粗体标注，底部还整整齐齐列着一排来源链 接。看上去就像一篇训练有素的记者写的快讯。 德国 DW 的分析团队在评估报告里写道：所有聊天机器人都用清晰易读的风格提供答案，这会让你产 ...

整理 | 褚杏娟 Sulaiman Ghori 在一期播客中，用了一个多小时详细讲述了他在 xAI 的经历。他说，在那里"从来没有人对我说不"，每个人都 被充分信任去做正确的事；只要是好想法，当天就能落地、当天就能得到反馈。他还提到，马斯克愿意被证明是错的，只要你 能拿出实验数据。 他也坦言，在上一家公司，很多事情也许他一个人能做得更快；但在 xAI，整体反而更快，因为几乎没有官僚流程。这些话， 听起来都是对公司的认同和马斯克的赞扬，实际上他还说自己是马斯克粉丝。 然后，播客发出来后第 3 天，他被解雇了。 外界猜测是因为他说了太多敏感信息。节目中，他透露了利用闲置特斯拉汽车驱动的人类模拟器 AI 代理的计划、还有马斯克 如何快速构建 Colossus 超级集群、xAI 在模型策略上的核心决策，曝光了公司内部部署测试的 AI 虚拟员工等，还有 xAI 也被 完全曝光。他坦率地谈到了激进的时间表、马斯克亲自参与的 Cybertruck 奖金计划、内部文化和运营方式以及一些非公开的 策略，这些言论引发了外界的强烈反响。 Sulaiman 自 2019 年起持续创业。在德国上大学一个月后退学，为了实现童年创办航天公 ...

Core Viewpoint - The article discusses the evolving landscape of AI tools, highlighting the lack of user loyalty and the rapid changes in preferences among users as new tools emerge and existing ones improve [5][14][39]. Group 1: AI Tools and User Behavior - AI tools are experiencing a surge in development, with significant advancements expected by 2025, leading to a competitive environment where users frequently switch between tools based on their immediate needs [8][9]. - Users exhibit a "cyber infidelity" behavior, quickly moving from one AI tool to another based on performance and specific functionalities, rather than maintaining loyalty to a single tool [14][16]. - The article illustrates the author's experience with various AI tools, emphasizing the importance of reliable information and the ability to adapt to changing requirements [16][18][20]. Group 2: Market Dynamics and Trends - The launch of Gemini3 has significantly impacted the market, with its capabilities leading to a rapid increase in demand and price for access, demonstrating the volatility and potential profitability in the AI tool market [30][34]. - The article notes that the introduction of new AI tools can disrupt existing user habits, prompting users to reconsider their tool choices and subscription models, such as preferring monthly over annual subscriptions to remain flexible [36][37]. - The competitive landscape is characterized by a constant influx of new tools, which forces users and businesses to evaluate the longevity and utility of each tool, impacting their purchasing decisions [36][40]. Group 3: Ecosystem and Integration - The article highlights the shift towards integrated ecosystems, where users find themselves relying on a suite of tools from a single provider, such as Google's ecosystem, due to its comprehensive capabilities [39][43]. - The need for seamless coordination between different AI tools is emphasized, with users expressing frustration over the lack of multi-modal integration and the challenges of switching between various platforms [45][50]. - The future of AI tools is anticipated to focus on unifying multiple models into a single interface, enhancing user experience and operational efficiency [50].

Core Insights - The article discusses the evolving landscape of AI tools, highlighting the lack of user loyalty and the rapid adoption of various AI models as they emerge in 2025 [12][13][46]. - It emphasizes the competitive nature of AI tools, where users frequently switch between different platforms based on their immediate needs and functionalities [10][15][46]. Group 1: User Behavior and Preferences - Users exhibit a "cyber infidelity" attitude towards AI tools, constantly switching based on which tool meets their current needs best [10][12]. - The emergence of new AI tools like Gemini3 has led to a significant shift in user preferences, with users quickly abandoning previous favorites for newer, more capable options [22][34]. - The article illustrates how users prioritize functionality and reliability over brand loyalty, often choosing tools that save them time or enhance their productivity [15][19][46]. Group 2: AI Tool Performance and Features - Gemini3 is highlighted for its impressive ability to analyze YouTube videos, outperforming other tools in terms of accuracy and information density [22][34]. - Users have reported that while tools like ChatGPT have improved, they still face challenges with context retention and accuracy, leading to frustration [20][52]. - The article notes that the integration of AI tools into existing workflows is becoming more seamless, with users increasingly relying on comprehensive ecosystems like Google's suite of tools for their functionality [44][46]. Group 3: Market Dynamics and Trends - The rapid rise in demand for AI tools has led to a booming market for account reselling, with prices for access to tools like Gemini3 skyrocketing shortly after their release [30][34]. - The article suggests that as new AI tools emerge, the market will continue to evolve, with users advised to opt for flexible subscription models to adapt to changing preferences [37][46]. - The future of AI tools is expected to focus on multi-modal capabilities and seamless integration across different platforms, as users seek more cohesive experiences [49][55].

Core Insights - The year 2025 is marked as a transformative period for artificial intelligence (AI), transitioning from experimental concepts to practical applications in everyday life, signifying the "Year of AI for All" [2] Group 1: Major Developments in AI - DeepSeek launched its R1 model on January 20, 2025, creating a significant impact in the global AI community [4] - On February 17, xAI released its latest AI model, Grok 3, followed by the introduction of the Evo 2 AI biology model by a collaboration of institutions on February 19 [6] - Google unveiled its strongest reasoning model, Gemini 2.5 Pro, on March 25, the same day OpenAI introduced a native image generation feature based on the GPT-4o model [6] Group 2: Notable Events and Milestones - The first humanoid robot marathon was held in Beijing on April 13, with the humanoid robot "Tiangong" winning the race [6] - By July 2025, China had registered 433 large models that were operational [8] - The EU's AI Act came into effect in August, establishing the first comprehensive regulatory framework for AI globally [8] Group 3: Product Launches and Innovations - On May 22, Anthropic launched the Claude 4 series of large models, and on July 18, OpenAI released a new intelligent product called "ChatGPT Agent" [8] - In October, ByteDance's Volcano Engine released an updated version of the Doubao model, which became the first in China to support "tiered adjustment of thinking length" [10] - Google's "Solar Catcher" project was publicly announced on November 5, aiming to create space-based machine learning data centers [12]

AI Model Development Roadmap - Grok 3 was released in February [1] - Grok 4 was released in July [1] - Grok Imagine was released in July [1] - Grok Code Fast 1 was released in August [1] - Grok 4 Fast was released in September [1] - Grokipedia was released in October [1] - Grok 4.1 was released in November [1] - Grok 4.1 Fast was released in November [1] - Grok Voice Agent API is scheduled for release in December [1]

RT Testlabor (@testerlabor)Did you know that Grok 3, Grok 4 and Grok 4.1 were all released within one year? https://t.co/jBFi0bUMf2 ...

Core Insights - A new AI startup, Humans&, is seeking to raise $1 billion with a target valuation of $4 billion, founded by Eric Zelikman, a former researcher at xAI [2][12] - Zelikman aims to develop AI models that learn user behavior and empathize with users, addressing the limitations of current reinforcement learning paradigms [2][17] - The startup's mission is to create AI that better understands human goals and emotions, moving beyond traditional task-oriented models [12][20] Company Overview - Humans& was co-founded by Eric Zelikman, who previously worked at xAI and contributed to the development of significant AI models [4][6] - The company is currently recruiting technical staff, offering competitive salaries starting at $350,000 annually [18] Technology and Innovation - Zelikman has developed the STaR algorithm, which enhances language models' reasoning capabilities, and has been recognized in top AI conferences [11][12] - The focus of Humans& is on creating AI that can collaborate with humans and understand diverse human aspirations and values [17][20] Market Context - The AI industry is shifting towards models that not only possess high intelligence but also emotional intelligence, reflecting a growing demand for more human-like interactions [20]