作者 | 王启隆 出品丨AI 科技大本营（ID：rgznai100） 2026 年 1 月，硅谷的风有点冷。 65 岁的 Yann LeCun（杨立昆）离职 Meta 创业还没一个月，就炮轰起了他的老东家，和那位他看不起的小将 Alexandr Wang。 虽然 LeCun 离开的新闻已经有一段时间，但如果你从 2013 年就开始关注这家公司，关注这群人，你会明白，这事实际上宣告了一个时代的彻底终 结。那个大厂愿意花重金养着一群科学家，让他们在一个叫 FAIR（Facebook AI Research）的象牙塔里，不计成本、不问产出、只为探索"智能本 质"的时代， 结束了 。 取而代之的，是 Alexandr Wang 这种年轻狠角色的上位，是"暴力美学"，是"别跟我谈哲学，把算力堆上去"，是赤裸裸的商业变现。 回首 2025 年，咱们公众号的爆款文章就献给了 LeCun。当时他在英伟达 GTC 大会意气风发，怼天怼地： 杨立昆"砸场"英伟达：不太认同黄仁勋，目 前大模型的推理方式根本是错的，token 不是表示物理世界的正确方式｜GTC 2025 2026 年开篇，我们聊聊这个倔强的法国人，是怎么把 Met ...

Alphabet could be a huge AI growth story over the coming years.While Alphabet (GOOGL +0.01%) (GOOG 0.18%) has been the best performer among the so-called "Magnificent Seven" stocks this year, it still may be one of the most underrated artificial intelligence (AI) stocks around and has one of the best growth stories both next year and over the long term.Let's look at why Alphabet is a stock you'll want to invest in.Controlling the entire tech stackWhen OpenAI introduced AI chatbots to the masses, Alphabet's ...

Core Viewpoint - The article discusses the rapid development and deployment of Google's Tensor Processing Unit (TPU), highlighting its significance in deep learning and machine learning applications, and how it has evolved to become a critical infrastructure for Google's AI projects [4][5][10]. Group 1: TPU Development and Impact - Google developed the TPU in just 15 months, showcasing the company's ability to transform research into practical applications quickly [4][42]. - The TPU has become essential for various Google services, including search, translation, and advanced AI projects like AlphaGo [5][49]. - The TPU's architecture is based on the concept of systolic arrays, which allows for efficient matrix operations, crucial for deep learning tasks [50][31]. Group 2: Historical Context and Evolution - Google's interest in machine learning began in the early 2000s, leading to significant investments in deep learning technologies [10][11]. - The Google Brain project, initiated in 2011, aimed to leverage distributed computing for deep neural networks, marking a shift towards specialized hardware like the TPU [13][15]. - The reliance on general-purpose CPUs for deep learning tasks led to performance bottlenecks, prompting the need for dedicated accelerators [18][24]. Group 3: TPU Architecture and Performance - TPU v1 was designed for inference tasks, achieving significant performance improvements over traditional CPUs and GPUs, with a 15x to 30x speedup in inference tasks [79]. - The TPU v1 architecture includes a simple instruction set and is optimized for energy efficiency, providing a relative performance per watt that is 25 to 29 times better than GPUs [79][75]. - Subsequent TPU versions, such as TPU v2 and v3, introduced enhancements for both training and inference, including increased memory bandwidth and support for distributed training [95][96].

Core Insights - Alphabet Inc. is a major player in the technology sector, primarily known for its search engine Google, and has expanded into AI and cloud computing, competing with giants like Amazon and Microsoft [1] Group 1: Price Target and Stock Performance - Guggenheim set a price target of $375 for GOOGL, indicating a potential increase of about 17.12% from its trading price of $320.18 [2][6] - The current stock price reflects a slight increase of 0.23, or 0.07%, with a trading range today between $316.79 and $326.83 [2] Group 2: AI Strategy and Development - Alphabet's strategic focus on AI infrastructure has evolved over a decade, starting with Google Brain in 2011 and the development of the TensorFlow framework [3] - The acquisition of DeepMind in 2014 significantly enhanced Alphabet's AI capabilities, culminating in the 2023 merger of Google Brain and DeepMind to develop the Gemini LLM [3] Group 3: Competitive Position and Market Capitalization - Alphabet's custom AI chips provide a significant cost advantage, enhancing its competitive edge in AI and cloud computing [4] - With a market capitalization of approximately $3.86 trillion, Alphabet is positioned as a must-own stock for investors interested in the future of AI [4][6] Group 4: Trading Volume and Stock Trends - Today's trading volume for GOOGL is 19.85 million shares, with the stock experiencing a high of $328.83 and a low of $140.53 over the past year [5] - Alphabet's sustained focus on innovation in AI infrastructure is expected to drive future growth and sector dominance [5]

Core Viewpoint - Alphabet is positioned as a leader in the AI infrastructure race, having developed its capabilities over the past decade, and is expected to widen its lead in the AI sector moving forward [2][9]. Group 1: AI Development and Infrastructure - Alphabet has been working on AI since 2011, establishing the Google Brain research lab and developing the TensorFlow framework, which is now widely used for training large language models (LLMs) [3][4]. - The company acquired DeepMind in 2014 and merged it with Google Brain in 2023, which contributed to the development of its Gemini LLM [3]. - Alphabet released its TensorFlow machine learning library in November 2015 and introduced tensor processing units (TPUs) in 2016, designed specifically for machine learning and AI workloads [4]. Group 2: Competitive Advantage - Alphabet's TPUs are now in their seventh generation, providing a significant performance and cost advantage over competitors who are just beginning to develop their own AI ASICs [5][6]. - The company benefits from a structural cost edge due to its combination of custom AI chips and foundational AI models, creating a flywheel effect that enhances its competitive position [6][7]. - By utilizing TPUs for training Gemini, Alphabet achieves better returns on capital expenditure compared to competitors relying on Nvidia's GPUs, allowing for reinvestment into further improvements [7]. Group 3: Market Position and Future Outlook - Alphabet's ownership of a world-class AI model enables it to capture the entire AI revenue stream, unlike competitors such as Amazon and Microsoft, which depend on third-party LLMs [8]. - The upcoming acquisition of cloud security company Wiz is expected to enhance Alphabet's ecosystem advantage [8]. - The company is positioned to be the big winner in the next phase of AI infrastructure due to its vertical integration and custom AI chips, making it a long-term buy despite its recent strong performance [9].

尽管OpenAI与英伟达主导了当前的人工智能叙事，但谷歌作为大模型核心架构Transformer的提出者， 以及曾经AlphaGo的开发方，其在AI领域的重要性一直不可忽视。近日，谷歌再次证明自己仍是赛场上 的关键力量，其最新一代大模型Gemini 3获业界高度评价，自主研发的TPU芯片更被传获科技巨头大规 模采购。 谷歌"模型+AI芯片"的组合，同时对OpenAI的软件优势与英伟达的硬件统治构成直接挑战。11月26日， 英伟达公开回应称："乐见谷歌的成功""英伟达技术依然领先行业一代"。不过，AI之战远未结束，业内 人士认为，随着日后产品的更新换代，AI时代的真正"霸主"仍有变数。 大模型芯片双管齐下 英伟达之所以公开回应，是因为近期有市场声音指出，该公司在AI基础设施领域的主导地位可能受到 谷歌芯片的威胁。当地时间11月25日，英伟达股价一度跌超6%，最终收盘跌幅缩小至2.59%。 本报（chinatimes.net.cn）记者石飞月 北京报道 智参智库特聘专家袁博对《华夏时报》记者表示，谷歌TPU是专为AI某些场景，如大模型的训练和推理 定制的AI芯片，最大特点是与谷歌自有的TensorFlow等工具以 ...

Core Viewpoint - Alphabet is positioned to be the biggest winner in the AI sector due to its structural cost advantages and vertical integration in AI technology [1][3]. Group 1: Market Position and Competitors - Nvidia currently dominates the GPU market for AI, while AMD is attempting to gain market share [2]. - Broadcom is assisting companies in developing custom ASICs for AI workloads, but Alphabet's internal development of AI chips gives it a competitive edge [2][5]. - Alphabet's Tensor Processing Units (TPUs) are in their seventh generation and optimized for its cloud infrastructure, providing a significant performance and energy efficiency advantage [5][6]. Group 2: Cost Efficiency and Revenue Opportunities - The shift from AI training to inference makes compute efficiency increasingly important, and Alphabet's TPUs consume less power, leading to lower operational costs [4][6]. - Alphabet does not sell its TPUs directly; instead, customers must use Google Cloud, allowing the company to capture multiple revenue streams within AI [7]. - By utilizing its own TPUs for internal AI workloads, Alphabet gains a cost advantage in developing and running its Gemini AI model compared to competitors relying on GPUs [8]. Group 3: Technological Advancements and Future Prospects - Alphabet's vertical integration and comprehensive AI tech stack position it favorably for future growth, with its Gemini 3 model receiving positive analyst reviews [9]. - The company's software platforms, such as Vertex AI, and its fiber network enhance its AI capabilities and reduce latency [10]. - The acquisition of cloud security company Wiz will further strengthen Alphabet's AI technology offerings [10].

Core Insights - The article highlights the inspiring journey of Soumith Chintala, who faced numerous rejections but ultimately created PyTorch, a significant tool in the AI landscape [1][10][22] Group 1: Background and Challenges - Soumith Chintala had a humble beginning, born in Hyderabad, India, and attended a second-tier university [2] - He faced significant challenges, including poor math skills and being rejected by 12 U.S. universities despite scoring 1420 on the GRE [4] - After obtaining a J-1 visa, he struggled to find direction and funding for further studies, leading to a series of rejections from graduate programs [4][5] Group 2: Career Development - Initially, Soumith worked as a test engineer at Amazon before joining Facebook AI Research (FAIR) [4][5] - He started as a low-level engineer but gained recognition after identifying and fixing a critical bug in an ImageNet task [5][6] - Despite initial skepticism about his project, he and his team decided to revamp Torch7, leading to the creation of PyTorch [8][9] Group 3: PyTorch's Impact - PyTorch was officially open-sourced in 2017 and quickly gained traction among top research labs, becoming a mainstream tool for deep learning [10][19] - The framework's flexibility and intuitive design allowed researchers to experiment more freely, leading to a rapid increase in its adoption [17][19] - By 2021, PyTorch's search volume surpassed that of TensorFlow, indicating its growing popularity in the AI community [17][21] Group 4: Community and Legacy - PyTorch has evolved from a niche framework to a foundational tool in AI, with a vast community of developers contributing to its ecosystem [21][26] - Soumith's journey from being rejected multiple times to becoming a respected figure in AI exemplifies resilience and dedication [22][27] - The framework is now integral to many leading AI models, including OpenAI's GPT series and Stability's generative models [26][30]

Core Insights - A unique team of two 14-year-old overseas students, LayersOfLogic, has gained attention at the 2025 Global "AI + Radio" Challenge, showcasing the potential of the younger generation [1][2] Group 1: Team Members - Victoria Wang, a Year 10 student at St Paul's Girls' School in the UK, excels in academics and extracurricular activities, including robotics and mathematics competitions, and demonstrates a well-rounded talent in sports and music [1] - Kevin Ke, a Year 10 student at Eton College, has a strong interest in biology, science, and mathematics, and is a music scholarship recipient, actively participating in various artistic and athletic activities [2] Group 2: Learning and Development - The team began with foundational knowledge in wireless communication and artificial intelligence, utilizing online tutorials to learn about IQ signals and signal preprocessing techniques [3] - They demonstrated mature teamwork skills, overcoming scheduling challenges through careful planning and communication, and learned the importance of perseverance in problem-solving [3] Group 3: Achievements and Future Aspirations - The experience of participating in the competition has significantly enhanced their knowledge and skills, allowing them to progress from basic Python to proficient use of TensorFlow for programming and data handling [3] - Both students expressed a desire to continue learning and exploring science and technology, applying the teamwork and problem-solving skills gained from the competition to future endeavors [4]

Core Insights - Soumith Chintala, the founder of PyTorch, announced his resignation from Meta after 11 years, marking a new leadership phase for the popular open-source deep learning framework [2][4] - PyTorch has become a core pillar in global AI research, supporting exascale AI training tasks and achieving over 90% adoption among major AI companies [2][9] Group 1: Chintala's Contributions and Career - Chintala played a pivotal role in advancing several groundbreaking projects at Meta's FAIR department, including GAN research and the development of PyTorch [5][12] - He rose from a software engineer to vice president in just eight years, a rapid ascent closely tied to the rise of PyTorch [5][10] - His departure comes amid significant layoffs at Meta AI, affecting around 600 positions, including those in the FAIR research department [4][6] Group 2: PyTorch's Development and Impact - PyTorch, created in 2016, evolved from the earlier Torch project and has become the standard framework in both academic and industrial settings [12][15] - The framework's success is attributed to its community-driven approach, user feedback, and the integration of features that meet real-world needs [15][16] - PyTorch has gained a reputation for its ease of use and flexibility, making it a preferred choice among researchers and developers [15][16] Group 3: Future Directions and Chintala's Next Steps - Chintala expressed a desire to explore new opportunities outside of Meta, emphasizing the importance of understanding the external world and returning to a state of "doing small things" [20][21] - He acknowledged the strong leadership team now in place at PyTorch, which gives him confidence in the framework's future [21]