Grok is far better than any AI agent account from last years craze. ...

Core Insights - Perplexity's CEO Aravind Srinivas argues that the deployment of AI agents relies on integrating them into real-world environments, specifically browsers, rather than solely enhancing model intelligence [1][2][4] - The Comet browser is positioned as a platform that allows AI agents to operate within familiar user contexts, leveraging existing user data and permissions [1][2][4] - Srinivas emphasizes that true intelligence is about executing tasks accurately rather than engaging in fanciful dialogue [2][4] Group 1 - The Comet browser is designed to be a "cognitive resource pool," allowing users to interact with AI agents in a seamless manner without compromising data privacy [2][4][10] - The browser's architecture is based on Chromium, which facilitates user migration and ensures a familiar interface, enhancing user adoption [8][9][10] - Comet aims to provide a natural and intuitive user experience, enabling users to delegate tasks to AI agents without the need for complex interactions [12][22][24] Group 2 - The current primary use case for Comet involves calling a sidecar assistant to complete tasks on the current webpage, showcasing its practical applications [12][24] - The company acknowledges that while Comet excels in certain tasks, it still struggles with complex, long-duration tasks that require multi-step coordination [25][26] - Srinivas believes that advancements in reasoning models will enable Comet to automate more complex workflows within the next six to twelve months [28][29] Group 3 - The company is focused on building a strong brand and user base, leveraging its existing Perplexity users to transition them to Comet [59][60] - Future plans include launching independent mobile apps for both Comet and Perplexity, catering to different user needs [54][66] - The company is considering partnerships and collaborations, including potential integration with Apple products, to enhance its market presence [64][66]

AI Agent Development - ChatGPT 发布了其自己的 AI Agent [1] - 该 AI Agent 使用一套完整的工具，包括可视化浏览器、文本浏览器、终端和直接 API [1] - AI Agent 能够计划旅行、分析数据和创建演示文稿 [1] Implications - 每个人现在都拥有了自己的 EA (Expert Assistant) [1]

Group 1 - The core viewpoint of the news highlights the performance of the Huaxia Sci-Tech AI ETF (589010), which has shown a slight increase of 0.10% after an initial surge and subsequent pullback, indicating a stable trading environment [1] - Among the holdings, Foxit Software led with a gain of 5.08%, while companies like Hehe Information, Yuke Technology, and Kingsoft Office also saw increases of over 3%. Conversely, Stone Technology experienced a decline of 2.14%, with other companies like Star Ring Technology, Lexin Technology, and Chip Origin also falling by more than 1% [1] - The trading volume reached a turnover rate of 15.53%, with a total transaction amount of 11.64 million, reflecting a high-frequency trading style in the market [1] Group 2 - MyShell, an AI agent platform, announced a significant update to its flagship product, which is described as the world's first AI product integrating "Claude Code + Cursor + v0" and other possibilities, likened to a combination of popular applications like "Jianying + Douyin" or an AI agent version of Instagram [1] - Guolian Minsheng Securities believes that China possesses mature technology for developing AI agents, with the market potential expected to gradually be released [1] - The year 2025 is projected to be a pivotal year for the growth of AI glasses, with significant movements from domestic and international manufacturers, including the official launch of Xiaomi's AI glasses, marking the entry of the consumer-grade smart glasses market into a phase of scaled competition [1][2]

AI Agent Capabilities - AI agents can automate tasks like compiling data into spreadsheets, potentially saving 4-8 hours of work [2] - AI agents can source information from the web, access PDF files, and extract data [3] - AI agents can generate spreadsheets with a high degree of accuracy, achieving up to 98% accuracy in some cases [4] - AI agents can format Excel workbooks according to instructions [4] - AI agents can perform 90-95% of time-consuming tasks, significantly reducing workload [5] Use Case: Budget Analysis - A specific use case involves analyzing the city of San Francisco's annual budget expenses and revenues for the past 5 years [2] - The AI agent can extract approximately 200 numbers from each PDF budget document [4] Efficiency and Time Savings - Using AI agents can significantly reduce the time spent on tasks, allowing users to focus on other activities [3][5] - Revisions can be made either within Excel or directly in the chat interface [4]

AI Agent Development - ChatGPT 发布了其自己的 AI Agent [1] - 该 AI Agent 使用一套完整的工具，包括可视化浏览器、文本浏览器、终端和直接 API [1] - 该 AI Agent 能够计划旅行、分析数据和创建演示文稿 [1] Implications for Users - 现在每个人都拥有了自己的 EA (AI Agent) [1]

AI Agent Development - ChatGPT released its own AI agent [1] - The AI agent utilizes a full suite of tools, enabling trip planning, data analysis, and presentation creation [1] - The AI agent includes a visual browser, a text browser, a terminal, and direct APIs [1] Technological Implications - The release of ChatGPT's AI agent means everyone now has their own EA (Executive Assistant) [1]

OpenAI is adding new AI agent features to ChatGPT https://t.co/WkLEfGuMMj ...

henry 发自 凹非寺 量子位 | 公众号 QbitAI 报告指出：在软件开发、数学竞赛、科学问答等任务中，agent已能完成 相当于人类花费50–200分钟才能完成的任务 ，并且这种能力还在快 速提升——大约每 2–6个月 就能 翻一番 。 在计算机操作任务中，虽然任务时长较短，但增长率与软件开发等任务一致。 Agent在自动驾驶任务的性能增长速度则较慢，约20个月翻一番。 Agent能力每7个月翻一番！ 根据非营利研究机构METR最新发布的报告，这一规律已在9项基准测试中得到了验证。 这些任务涉及编程、数学、计算机使用、自动驾驶等领域，表明大模型正在不断向着高度自动化迈进。 在视频理解任务中，模型能够在 时长1小时 的视频上取得 50% 的成功率。 作为一家致力于研究前沿人工智能系统能力及其风险的研究团队，METR此次的报告又进一步拉近了AI自主化的时间线，快来和我们看看报告 有哪些内容吧。 Agent的摩尔定律 在此前的测试中，METR将评估范围聚焦于软件开发和研究类任务，并发现AI agent的能力呈现出一种"摩尔定律"式的增长趋势—— 平均每七 个月，其可完成任务的time horizon就会翻一 ...

Summary of Key Points from Conference Call Records Industry Overview - The conference call discusses the progress of AI applications in both B-end and C-end markets globally and in China, highlighting the rapid development and commercialization of AI technologies in various sectors [1][2][3]. Core Insights and Arguments - **C-end AI Applications**: China leads in C-end AI applications, benefiting from a large user base and mobile internet advantages. Companies like ByteDance, Kuaishou, and Meitu have shown significant AI integration in their products, reflected in their financial results [2]. - **B-end AI Applications**: The B-end market in China faces challenges such as low willingness to pay for AI solutions. However, sectors like manufacturing, education, and healthcare are leveraging AI to enhance competitiveness and provide customized experiences [1][3][5][6]. - **Meitu's Growth**: Meitu has seen a significant increase in paid users, reaching 12.61 million in 2024, more than doubling from 2022, due to its user-friendly AI product designs [10][11]. - **Applovin's XOY AI Engine**: Applovin's XOY AI advertising engine has driven performance growth by transitioning from a click-based payment model to a performance-based model, enhancing ad targeting and user engagement [13]. - **Tems AI's Business Model**: Tems AI has achieved commercial success through precision medicine, generating $445 million in revenue in 2024 by charging hospitals per case and earning commissions from pharmaceutical companies [15]. Important but Overlooked Content - **Challenges in B-end Market**: The B-end market in China struggles with low payment rates compared to the U.S., where companies are more accustomed to paying high premiums for software. This necessitates a focus on demonstrating the tangible benefits of AI solutions [12]. - **AI in Manufacturing**: AI applications in manufacturing are expected to enhance production efficiency and quality, with companies like Dingjie Digital leading the way in implementing AI tools [5][22][24]. - **Growth of AIGC in C-end**: The C-end generative AI applications have rapidly grown, with apps like Deepseek and Doubao surpassing 100 million monthly active users, indicating a strong market potential [7][8]. - **Investment Opportunities**: Companies in enterprise service software, such as Kingdee International and Yonyou Network, as well as those in vertical industries like healthcare and education, are highlighted as having strong growth prospects [30]. Conclusion - The conference call emphasizes the transformative impact of AI across various sectors, with notable advancements in both B-end and C-end applications. The potential for growth in AI technologies remains significant, particularly in the context of China's unique market dynamics and user base.