AI 与公司行政工作相结合具有一定的挑战性，但也存在可能的方向。目前大多数的“AI 应用/AI 转型”在行政工作方面还在走“数字化转型”的老路，把 AI 往现有流程上套，讲“固化流程”“节约成本”的故事。但在技术加速迭代的当下，这样做可能导致成果过时，剥夺企业主动进化的能力。
YCombinator 的圆桌讨论认为，垂直 AI 智能体的市场潜力巨大，其专注于特定领域，能提供定制化服务并自动化重复任务，从而提高效率和降低成本。创业者应关注行政任务领域，这或许能为 AI 与行政工作的结合提供思路。
然而，AI 并非万能，我们和 AGI 还差得很远。不能仅因对 AI 的焦虑就希望其拿来即用、马上见效。AI 的力量不应只用于现有业务流程的优化，而更应用于对未来业务的重新定义。比如像电力发明时，不应从“如何让电力赋能马车”出发，而应从“电力能创造和满足什么新的需求”出发。
现在的AI不仅仅是流量密码，也是股价密码；于是，是个公司都会想方设法往AI上沾边。私下也有很多朋友问我：美妆/白酒/奢侈品如何联动AI？AI如何赋能农业/传统制造业?AI如何赋能HR/行政/采购/公关？……其实，大多数都有点难。因为AI也并不是万能的，以及我们和AGI还差的很远。像我去年12月份的文章[《AI原生公司|未来打工人》](http://mp.weixin.qq.com/s?__biz=MzkyMTY1MTM4Mw==&mid=2247483844&idx=1&sn=3adfc54f294c53fdf51105cdb4e23ec3&chksm=c181101cf6f6990af5116997115fdfcf95360bb55a14f05ea9aed00239c88dfb440d03055533&scene=21#wechat_redirect)写的，目前大多数的"AI应用/AI转型”还在走“数字化转型”的老路：把AI往现有流程上一套，还在讲“固化流程”、“节约成本”的故事。但在技术加速迭代的今天，这样做基本就等于“做出来就是过时的”：把企业的业务模式凝固在今天，同时又剥夺了企业主动进化的能力。现在这种情况下反映出来的，更多是人们对于AI的焦虑：所以才会希望AI拿来就能用，马上能起效果。但我们不能止于焦虑：AI的力量并不应该只用在现有业务流程的优化上，而更应该用在对于未来业务的重新定义之上。这才是“AI原生公司”应该的做法。就像在电力发明的时候，我们不应该从“如何让电力赋能马车”出发，而应该从“电力能创造和满足什么新的需求”出发。
The government will complement its context-specific approach to AI regulation by proposing a proportionate‘layered approach’to applying available AI technical standards.This involves regulators identifying relevant technical standards and encouraging their adoption by actors in the AI life cycle to support the integration of the AI regulation principles into technical and operational business processes:Layer 1:To provide consistency and common foundations across regulatory remits,in the first instance regulators could seek to encourage adoption of sector-agnostic standards which can be applied across AI use cases to support the implementation of cross-sectoral principles.For example,management systems,risk management,and quality standards161 can provide industry with good practices for the responsible development of AI systems.The adoption of these standards should be encouraged by multiple regulators as tools for regulated entities to establish common good practices for AI governance.Layer 2:To adapt these governance practices to the specific risks raised by AI in a particular context,regulators could look at encouraging adoption of additional standards157 AI-specific standards addressing trustworthiness characteristics such as safety,transparency and robustness,amongst others,have been developed or are currently being developed(“*”indicates standards which are under development at the time of writing)in SDOs such as ISO/IEC and IEEE(e.g.IEEE 7001,ISO/IEC TS 6254*,ISO/IEC TR 5469*,ISO/IEC 240292*).158 Technical standards can be updated as good practices and the technology develop,allowing flexibility for requirements to adapt to technological change.159 Standards help organisations to manage and mitigate risks,as well as helping to unlock and scale the benefits of their products and services.In doing so,standards play a role in responsible innovation both as tools supporting good governance and as mechanisms for enabling and accelerating innovation.
《[YC圆桌：垂直AI智能体的规模可能是SaaS的十倍](https://mp.weixin.qq.com/s/jx_2-2jji8MuGu7keU3YDQ)》YCombinator的圆桌讨论认为，垂直AI智能体的市场潜力可能是SaaS的十倍。垂直AI智能体专注于特定领域，提供定制化服务并自动化重复任务，从而提高效率和降低成本。与SaaS相比，AI智能体能直接取代人工操作，推动企业管理效率提升和运营模式变革。创业者应关注无聊的行政任务领域，以便捕捉市场需求。《[甲小姐对话张钹：中国大模型的死与生](https://mp.weixin.qq.com/s/v_R2EYPnrlLQzwSXqIpN7w)》张钹院士指出，中国大模型的发展面临资金和市场的双重挑战，强调必须与应用相结合才能生存。他认为，美国大模型企业能够专注于提升质量，而中国则需要寻找商业闭环，避免仅依赖投资。张钹建议企业优先考虑盈利，警惕过度追求技术而忽视市场需求，以确保在竞争激烈的环境中生存。《[Robotaxi 2024年度格局报告](https://waytoagi.feishu.cn/record/JQreri9meeUrteczb42cqT3Dn6Z)》由量子位发布，分析了Robotaxi的市场现状、竞争要素和未来趋势。报告指出，技术、政策和商业化运营是影响Robotaxi落地的关键因素，其中技术是核心驱动力。全球多个城市已开始Robotaxi的运营，中国市场尤为活跃。报告预测，Robotaxi市场规模将快速增长，2030年在网约车市场的渗透率接近50%。百度、小马智行、文远知行等公司在技术实力、运营积累和商业化进度上领先。特斯拉和Wayve等公司也在积极布局，推动行业发展。