编辑 | 硬 AI 要点提炼： "如果有什么东西出现了短缺，人类的历史证明它最终会被大规模复制，直到变成过剩——AI芯片和算力也不会例外。"a16z联合创始人预言，AI是比互联网更宏大的技术变革，目前 尚处于早期阶段。他指出，智能成本正以超摩尔定律的速度崩塌，基础设施的短缺终将导致产能过剩，从而利好应用层。 作者 | 硬 AI 成本"通缩" ： 智能的单位成本（Per unit cost）下降速度远超摩尔定律，这将带来需求的爆发性增长。 "短缺导致过剩" ： 遵循"短缺导致过剩"的历史规律，GPU和数据中心的大规模建设最终将导致供应过剩，从而进一步压低AI成本。 市场结构 ： 未来AI市场将类似计算机产业结构：少数"上帝级模型"（类似超级计算机）在顶端，海量低成本"小模型"在边缘侧普及。 中美竞争 ： 这是一个双雄争霸的局面。中国（如DeepSeek、Kimi）在追赶速度、开源策略和芯片自研上表现惊人，这促使美国联邦层面的监管风向转为支持创 新。他说道，"基本上AI只在美国和中国被构建。世界其他地方要么造不出来，要么不想造"。 商业模式 ： AI应用正从"按Token付费"向"基于价值定价"转移；初创公司不再 ...

大家好欢迎收看【古今财鉴】 美国这几年对中国AI的围堵，动作可不小。 又是高端芯片禁运，又是闭源模型技术壁垒，摆明了想卡住咱们的脖子。 财经摆渡人 精研出品 破浪前行，共探财富新局 当时不少人觉得，中国AI发展怕是要慢下来了。 谁能想到，现在风向变了超过一半的美国初创企业，居然把中国开源AI模型当成了开发首选。 这剧情反转得，连我这个常年追科技圈的都觉得意外。 就拿硅谷那些明星公司来说，不是偷偷用，是明着"站队"。 Perplexity这种AI创业独角兽，底层技术直接搭在中国模型上；Airbnb更直接，CEO公开说他们的AI客 服系统"重度依赖"阿里的Qwen模型。 我翻了翻硅谷工程师的聊天记录，DeepSeek、GLM、Kimi这些名字，现在跟咖啡一样成了办公室标 配。 不光是商业公司，连学术界和芯片巨头都"真香"了。 斯坦福大学搞推理模型研究，直接拿Qwen当基座改；英伟达训练自家AI时，居然用Qwen生成高质 量"合成数据"。 这场景挺讽刺的美国想卡中国的算力脖子，结果中国模型反倒成了美国AI的"养料供给者"。 半数美企"用脚投票"，中国模型成硅谷新宠 为什么美国公司放着自家的闭源模型不用，非要千里迢迢 ...

Core Viewpoint - The article emphasizes the importance of integrating artificial intelligence (AI) into various industries, particularly in the financial sector, to enhance service quality and operational efficiency while ensuring inclusivity and security in AI applications [3]. Group 1: AI Models - The choice between open-source and closed-source models is not just a technical issue but has profound implications for application. Open-source models promote equality and cost savings but may have slower iteration rates and higher error rates, while closed-source models offer stability and reliability but limit customization and transparency [4]. - The financial industry should focus on "AI+" rather than solely on building large models, combining the advantages of both open-source and closed-source models to enhance service quality and internal management efficiency [4]. Group 2: Decision-making AI vs. Generative AI - Decision-making AI excels in scenarios requiring high interpretability and accuracy, dominating over 80% of current applications in finance, particularly in risk assessment and fraud detection. In contrast, generative AI is more suited for creative tasks and is primarily used in non-core areas like customer service [5]. - The trend indicates that as the capabilities of large models improve, generative AI may see exponential growth and work in tandem with decision-making AI, blurring the lines between the two [5]. Group 3: AI Inclusivity and Computing Power - The demand for GPU computing power is expected to remain in a "tight balance" as AI becomes more widespread, necessitating efforts to optimize existing resources and expand capacity [8]. - Companies should adopt engineering methods to reduce operational costs and enhance resource efficiency while building high-performance computing centers to support AI applications [8]. Group 4: Safety and Security in AI Applications - As AI inclusivity increases, the stability and security of AI applications must be prioritized to protect public interests. This includes establishing safety measures and enhancing data quality to build trust in AI systems [9]. - There is a need to prevent model resonance to mitigate systemic risks, as the concentration of mainstream models may lead to vulnerabilities across institutions. Developing a reliable knowledge base and differentiated model training is essential for enhancing the resilience of the financial system [9].

今年夏天开始，扎克伯格在Meta砸重金打造"超级智能实验室"，对内部AI团队进行大重组，到处用高薪挖人，搞得硅谷同行人心惶惶。 但几个月过去，Meta尚未拿出实际成果。这在某种程度上情有可原，毕竟新模型的构建需要时间。多方信源显示，Meta目前的一个主线任务是在明年一 季度发布一个全新的闭源模型，这个项目内部代号"牛油果"。 但如今，最新爆料却显示，主导"牛油果"的Meta首席AI官Alexandr Wang也和扎克伯格合作得不愉快。 看起来，扎克伯格正在面临管理难题——在全力push的扎克伯格手下，似乎没有人干活顺心。杨立昆官宣离职、高层洗牌，就连新的AI负责人也怨声载 道。 扎克伯格摇着美金，招揽了诸多人才，并在内部给予他们高位。钱到位了，人也到位了，但这不代表着一切一定会朝着扎克伯格希望的方向顺利发展。 2026年的脚步近了，扎克伯格的AI梦却不一定成真。 弗里德曼也面临着扎克伯格日益增加的压力。扎克伯格总是要求他更快地交付AI产品，但这往往导致为了发布而发布的局面，出力却不讨好。 弗里德曼团队的一些成员感到非常沮丧，因为他们认为Vibes的发布过于仓促，以惊人的速度推出，只是为了抢在OpenAI类似产 ...

Investment Rating - The industry investment rating is "Outperform the Market" (maintained) [4][10] Core Viewpoints - The AI competition landscape is deepening in multiple dimensions, accelerating the evolution of ecological applications [2] - AI creative and reasoning applications are entering a high growth phase, with programming and role-playing becoming core scenarios [7] - The report highlights the significant user growth and engagement in the AI creative sector, with active users exceeding one million for leading products [7] Summary by Relevant Sections Industry Performance - The relative return over 1 month is -0.86%, over 3 months is -7.8%, and over 12 months is -4.17% [6] - The absolute return over 1 month is -3.01%, over 3 months is -7.71%, and over 12 months is 9.95% [6] AI Creative Sector Insights - In November 2025, over 200 AI creative web products had a total visit count exceeding 27 million, with independent visitors surpassing 7.7 million [7] - Leading products like Jimeng AI and Gaoding AI have over one million active users, with significant growth in independent visitor numbers for Canva and Keling AI [7] - The average usage time for several products exceeds 8 minutes, indicating strong user engagement [7] AI Development Trends - The AI development is entering a "great divergence" phase, with open-source models capturing 30% of traffic globally, and Chinese open-source models leading in specific scenarios [7] - The report notes a shift from generating text to problem-solving capabilities in AI, with programming requests increasing from 11% to over 50% [7] - The multi-model ecosystem is becoming clearer, with closed-source models handling high-value tasks and open-source models focusing on low-cost, high-concurrency needs [7] Investment Recommendations - The report suggests focusing on companies such as BlueFocus, Kunlun Wanwei, Tianyu Digital Science, Yinsai Group, Visual China, and HuiLiang Technology as potential investment opportunities [7]

（文/陈济深 编辑/张广凯） 12月10日，彭博社爆料称，扎克伯格组成了一个名为TBD Lab的团队。该团队在其最新模型"牛油 果"（Avocado）的训练过程中使用了多个第三方模型，包括谷歌的Gemma、OpenAI的GPT-oss和阿里巴 巴的Qwen模型。该款模型预计将于明年春季首次亮相，并可能作为"闭源"模型推出。 对此，Meta 的一位发言人则对外宣称："我们的模型训练工作正按计划进行，时间表没有发生有意义的 变更。" 消息被爆出后，阿里巴巴美股盘前一度上涨4%，收盘涨幅2.53%。 扎克伯格挥舞重金招来的AI大牛们计划开发的闭源大模型，竟然是通过中国的开源模型来训练，这不 仅意味着如今中国开源阵营的崛起，也代表扎克伯格曾经的美国开源霸主豪言，终究没有抵过来自中国 的竞争压力。 过气的开源盟主 Meta急着抄作业源于其开源旗舰模型Llama 4的失败表现。 过去两年，Meta通过开源Llama系列，成功扮演了"反OpenAI联盟"的盟主。Llama被视为开源界的 Linux，一度是全球开发者（包括中国开发者）的首选底座。 然而，这一格局在2025年开始瓦解。 随着年初DeepSeek开源模型的横空出 ...

阿里巴巴美股收盘上涨1.83%至158.82美元。 有消息称扎克伯格组成了一个名为TBD Lab的团队。该团队在"Avocado"的训练过程中使用了多个第三 方模型，包括谷歌的Gemma、OpenAI的GPT-oss和阿里巴巴的Qwen模型。新模型"Avocado"预计将于明 年春季首次亮相，并可能作为"闭源"模型推出。 ...

Group 1 - The core point of the article is that Mark Zuckerberg has formed a team called TBD Lab, which is working on a new model named "Avocado" that is expected to debut in spring next year [1] - The TBD Lab team is utilizing multiple third-party models during the training of "Avocado," including Google's Gemma, OpenAI's GPT-oss, and Alibaba's Qwen model [1] - Alibaba's stock price increased by 1.83% to $158.82 at the close of trading in the US [1]

有消息称扎克伯格组成了一个名为TBD Lab的团队。该团队在"Avocado"的训练过程中使用了多个第 三方模型，包括谷歌的Gemma、OpenAI的GPT-oss和阿里巴巴的Qwen模型。新模型"Avocado"预 计将于明年春季首次亮相，并可能作为"闭源"模型推出。 阿里巴巴美股收盘上涨1.83%至158.82美元。 编辑 ｜心一 记者 ｜ 一财科技 ...

男生普通话版 下载mp3 女声普通话版 下载mp3 粤语版 下载mp3 西南方言版 下载mp3 东北话版 下载mp3 上海话版 下载mp3 今日优选 美联储：降息25个基点，在未来30天内购买400亿美元短债 特朗普批评鲍威尔：降息幅度太小，可以翻倍，应降至全球最低 特朗普：美国在委内瑞拉附近扣押一艘委油轮，此举合法 加拿大央行如期按兵不动 美国第三季度劳动力成本增速降至四年最低 据悉Meta全力转向闭源模型 中国财政部发7500亿元特别国债"借新还旧"，不增加财政赤字 上期所调整白银期货相关合约交易保证金比例和涨跌停板幅度 市场盘点 周三，美联储如期降息25个基点，本月将进行400亿美元净资产购买。美元指数在剧烈波动中下跌，最终收跌0.60%，报98.65；基准的10年期美债收益率最 终收报4.1550%，对美联储政策利率敏感的2年期美债收益率收报3.5530%。 美联储宣布连续第三次降息后，现货黄金一度重回4230美元上方，最终收涨0.48%，报4228.55美元/盎司；现货白银日内大涨逾1美元，续创历史新高，最终 收涨1.89%，报61.81美元/盎司。 美国在委内瑞拉海岸扣押油轮，油价扭转跌势收盘走高 ...