凤凰网科技讯 北京时间11月26日，据《商业内幕》报道，谷歌公司最近势如破竹，在大模型和自研芯 片领域均拿出上佳表现，一下子成为了市场宠儿，让对手OpenAI、英伟达感受到了压力。 2022年底，当OpenAI凭借ChatGPT一飞冲天时，谷歌一度陷入困境。由于谷歌在推出自己的聊天机器人 过程中屡次失误，一些最关注谷歌的人士甚至呼吁CEO桑达尔·皮查伊(Sundar Pichai)辞职。 然而，接近三年后，谷歌完成了一次不可思议的逆转。该公司全新的AI模型Gemini 3被证明是一次成 功，以至于Salesforce CEO马克·贝尼奥夫(Marc Benioff)宣布他将从ChatGPT转向使用Gemini 3。谷歌的 市值刚刚超越了微软，并且正朝着4万亿美元俱乐部迈进。今年，谷歌母公司Alphabet股价上涨了近 70%。 这表明，一直拥有各项竞争资源的谷歌，终于让所有环节实现了协同运作，从AI模型到触达用户的搜 索引擎等平台。 在快速发展的AI领域，没有哪一次胜利是稳固的，但谷歌从未显得如此强大。以下是《商业内幕》总 结的五点原因： 1.Gemini 3"赶超"ChatGPT 的高增长公司。 谷歌在上周 ...

Core Insights - The article discusses the evolution of pre-training in AI, emphasizing its critical role in enhancing model performance through scaling laws and effective data utilization [5][8][9] - Nick Joseph, head of pre-training at Anthropic, shares insights on the challenges and strategies in AI model development, particularly focusing on computational resources and alignment with human goals [2][3][4] Pre-training Fundamentals - Pre-training is centered around minimizing the loss function, which is the primary objective in AI model training [5] - The concept of "scaling laws" indicates that increasing computational power, data volume, or model parameters leads to predictable improvements in model performance [9][26] Historical Context and Evolution - Joseph's background includes significant roles at Vicarious and OpenAI, where he contributed to AI safety and model scaling [2][3][7] - The transition from theoretical discussions on AI safety to practical applications in model training reflects the industry's maturation [6][7] Technical Challenges and Infrastructure - The article highlights the engineering challenges faced in distributed training, including optimizing hardware utilization and managing complex systems [12][18][28] - Early infrastructure at Anthropic was limited but evolved to support large-scale model training, leveraging cloud services for computational needs [16][17] Data Utilization and Quality - The availability of high-quality data remains a concern, with ongoing debates about data saturation and the potential for overfitting on AI-generated content [35][36][44] - Joseph emphasizes the importance of balancing data quality and quantity, noting that while data is abundant, its utility for training models is critical [35][37] Future Directions and Paradigm Shifts - The conversation touches on the potential for paradigm shifts in AI, particularly the integration of reinforcement learning and the need for innovative approaches to achieve general intelligence [62][63] - Joseph expresses concern over the emergence of difficult-to-diagnose bugs in complex systems, which could hinder progress in AI development [63][66] Collaboration and Team Dynamics - The collaborative nature of teams at Anthropic is highlighted, with a focus on integrating diverse expertise to tackle engineering challenges [67][68] - The article suggests that practical engineering skills are increasingly valued over purely theoretical knowledge in the AI field [68][69] Implications for Startups and Innovation - Opportunities for startups are identified in areas that can leverage advancements in AI models, particularly in practical applications that enhance user experience [76] - The need for solutions to improve chip reliability and team management is noted as a potential area for entrepreneurial ventures [77]

Core Insights - The discussion around the "AI bubble" has cooled down, with Deutsche Bank recommending a long-term investment strategy rather than attempting to time the market for optimal returns [1][13][19] Group 1: Investment Trends - Major tech companies are investing hundreds of billions in AI infrastructure, raising concerns about potential bubble risks [2][8] - OpenAI's CEO announced a $500 billion infrastructure plan called "Stargate," while Meta has committed to investing several hundred billion in data centers [2][11] - Bain & Company predicts that AI companies will need $2 trillion in annual revenue by 2030 to support required computing power, but actual revenue may fall short by $800 billion [1][2] Group 2: Market Sentiment - Deutsche Bank's research indicates that the search volume for "AI bubble" has significantly decreased, reflecting a typical pattern seen in previous market bubbles [13][15] - Concerns about AI investments are diminishing, with media sentiment dropping from 7.3 to 5.1 on a scale of 10 [13][15] Group 3: Financial Strategies - Deutsche Bank emphasizes the difficulty of accurately timing the market, citing historical examples where missing key trading days drastically reduced returns [17][19] - The bank advises investors to adopt a long-term holding strategy to capture the risk premium associated with equity investments [19][20] Group 4: Challenges in AI Development - AI technology faces challenges, including diminishing returns on increased computing power and data, as acknowledged by OpenAI's CEO [8][12] - A study from MIT found that 95% of organizations have not seen any returns on their AI investments [6][8]

Group 1 - Safe Superintelligence (SSI) announced personnel changes, with co-founder Daniel Gross leaving and Ilia Sutskever taking over as CEO [2] - Daniel Levy has been promoted to president of SSI following Gross's departure [2] - Gross has joined Meta as the head of the AI product division [2] Group 2 - SSI's valuation reached $32 billion after a funding round in April 2025, with investments from Alphabet and Nvidia [4] - Sutskever emphasized the need for a new research direction in safe superintelligence, diverging from his previous work at OpenAI [4] - Sutskever noted the limitations of data availability, stating, "We have reached the limits of data. After all, there is only one internet" [4] Group 3 - Meta is undergoing a significant AI recruitment drive, investing $14 billion in Scale AI to attract top talent [5] - The company has faced challenges, losing 11 of the original authors of the Llama research paper, which has exacerbated its technical difficulties [5] - Meta's investment strategy includes acquiring 49% of Scale AI to bring in its founder, Alexandr Wang, as a lab leader [5] Group 4 - The competition for talent between Meta and OpenAI has intensified, with OpenAI's CEO Sam Altman accusing Meta of offering large salaries to lure developers [6] - Meta's recruitment efforts include targeting reasoning experts to address its technical shortcomings [7] - An internal memo from OpenAI revealed concerns about the competitive landscape, indicating a sense of urgency in adjusting compensation strategies [7]