Institutional AI vs Individual AI
### 章节 1:引言:生产力去哪儿了? 📝 **本节摘要**: > 本章探讨了AI时代的一个悖论:尽管个人生产力因AI获得了显著提升,但企业整体价值却未见成比例的增长。作者通过1890年代工厂电气化革命的历史教训指出,仅仅引入先进技术而不去重塑组织架构,是无法带来实质性回报的。文章强调,高生产力...
Category: AI📝 本节摘要:
本章探讨了AI时代的一个悖论:尽管个人生产力因AI获得了显著提升,但企业整体价值却未见成比例的增长。作者通过1890年代工厂电气化革命的历史教训指出,仅仅引入先进技术而不去重塑组织架构,是无法带来实质性回报的。文章强调,高生产力的个体并不等同于高生产力的企业,真正的未来在于构建“制度级智能(Institutional Intelligence)”,即将技术与制度结合,打造未来的全新“流水线”。这也是接下来要论述的“制度级AI”与“个人AI”的核心分水岭。
[原文] [Author / George Sivulka]: Institutional AI vs Individual AI
[译文] [Author / George Sivulka]: 制度级AI vs 个人AI
[原文] [Author / George Sivulka]: Where did the productivity go?
[译文] [Author / George Sivulka]: 生产力去哪儿了?
[原文] [Author / George Sivulka]: AI just made every individual 10x more productive.
[译文] [Author / George Sivulka]: AI刚刚让每个个体的生产力提升了10倍。
[原文] [Author / George Sivulka]: No company became 10x more valuable as a result.
[译文] [Author / George Sivulka]: 但没有哪家公司因此变得具有10倍的价值。
[原文] [Author / George Sivulka]: Where did the productivity go?
[译文] [Author / George Sivulka]: 生产力去哪儿了?
[原文] [Author / George Sivulka]: This isn't the first time this has happened.
[译文] [Author / George Sivulka]: 这并不是历史上第一次发生这种情况。
[原文] [Author / George Sivulka]: In the 1890s, electricity promised enormous productivity gains.
[译文] [Author / George Sivulka]: 在1890年代,电力承诺了巨大的生产力增长。
[原文] [Author / George Sivulka]: Textile mills in New England, built to harness the rotational power of steam engines, quickly installed faster electric motors in their place.
[译文] [Author / George Sivulka]: 新英格兰的纺织厂原本是为利用蒸汽机的旋转动力而建的,它们迅速在原位安装了更快的电动机。
[原文] [Author / George Sivulka]: But for thirty years, electrified mills saw almost no increase in output.
[译文] [Author / George Sivulka]: 但在长达三十年的时间里,电气化的工厂几乎没有看到任何产量的增加。
[原文] [Author / George Sivulka]: The technology was far superior. But the organization was not.
[译文] [Author / George Sivulka]: 技术远胜从前。但组织方式却没有进步。
[原文] [Author / George Sivulka]: It wasn't until the 1920s, when factories completely redesigned the mills once again, with assembly lines, individual motors within every piece of equipment, and workers and machines executing drastically different jobs, that electrification produced meaningful returns.
[译文] [Author / George Sivulka]: 直到1920年代,当工厂以装配线、每台设备内置独立电机以及工人和机器执行截然不同的工作,彻底重新设计了这些磨坊时,电气化才产生了有意义的回报。
[原文] [Author / George Sivulka]: Fig. 1: The three evolutions of the Lowell Textile Mills. From left to right: the 1890 steam engine-powered mill, the 1900 electrical engine-powered mill, and finally the 1920 “unit drive” mill i.e. a ground up rebuild as an electrical assembly line.
[译文] [Author / George Sivulka]: 图1:洛厄尔纺织厂(Lowell Textile Mills)的三次演变。从左至右:1890年以蒸汽机为动力的工厂,1900年以电动机为动力的工厂,最后是1920年的“单元驱动(unit drive)”工厂,即作为一条电气化装配线彻底重建的工厂。
[原文] [Author / George Sivulka]: These returns came not from the technology itself, and not from making individual workers or machines faster at spinning thread.
[译文] [Author / George Sivulka]: 这些回报并非来自技术本身,也不是因为让个别工人或机器纺线纺得更快。
[原文] [Author / George Sivulka]: It was when we finally redesigned the institution and the technology together that the upside materialized.
[译文] [Author / George Sivulka]: 而是当我们最终将制度(institution)与技术结合起来重新设计时,上升空间才得以显现。
[原文] [Author / George Sivulka]: This is the most expensive lesson in the history of technology, and we're learning it again, right now.
[译文] [Author / George Sivulka]: 这是技术史上最昂贵的一课,而我们现在正重新学习它。
[原文] [Author / George Sivulka]: In 2026, AI is driving a 10x increase in the productivity of the individuals who know how to leverage it. But that's not enough.
[译文] [Author / George Sivulka]: 在2026年,AI正在推动那些知道如何利用它的个体的生产力实现10倍增长。但这还不够。
[原文] [Author / George Sivulka]: We've swapped the motor; we have not yet redesigned the factory.
[译文] [Author / George Sivulka]: 我们已经换掉了马达;我们尚未重新设计工厂。
[原文] [Author / George Sivulka]: Because of a simple fact: productive individuals do not make productive firms.
[译文] [Author / George Sivulka]: 因为一个简单的事实:高生产力的个体并不造就高生产力的公司。
[原文] [Author / George Sivulka]: The wide majority of AI products evoke the feeling of being productive, but they haven't moved the needle on driving value.
[译文] [Author / George Sivulka]: 绝大多数AI产品唤起了一种充满生产力的“感觉”,但它们在驱动价值方面并没有带来实质性的改变。
[原文] [Author / George Sivulka]: The majority of publicized AI use is individuals self-indulgently “productivity-maxxing” on Twitter or in company Slack channels, with zero real impact.
[译文] [Author / George Sivulka]: 大多数被公开宣传的AI使用案例,都是个人在Twitter或公司Slack频道里自我沉醉式地追求“生产力拉满(productivity-maxxing)”,产生着零实质影响。
[原文] [Author / George Sivulka]: The “services as software” motif that's been repeated for a year now points in the right direction, but offers no blueprint. And it misses the bigger picture.
[译文] [Author / George Sivulka]: 已经被重复了一年的“服务即软件(services as software)”主题指明了正确的方向,但没有提供蓝图。而且它忽略了更大的图景。
[原文] [Author / George Sivulka]: The real shift isn't from tools to services, it's building the technology and the institution together (whether legacy or new).
[译文] [Author / George Sivulka]: 真正的转变不是从工具到服务,而是将技术与制度(不论是传统体制还是新体制)一起构建。
[原文] [Author / George Sivulka]: A truly productive future requires an entirely new class of product. The assembly line of tomorrow.
[译文] [Author / George Sivulka]: 一个真正充满生产力的未来需要一类全新的产品。明日的装配线。
[原文] [Author / George Sivulka]: Productive organizations require “Institutional Intelligence.”
[译文] [Author / George Sivulka]: 高生产力的组织需要“制度级智能(Institutional Intelligence)”。
[原文] [Author / George Sivulka]: This essay will dive into the seven big factors that differentiate “Institutional AI” from “Individual AI.”
[译文] [Author / George Sivulka]: 本文将深入探讨区分“制度级AI(Institutional AI)”与“个人AI(Individual AI)”的七大关键因素。
[原文] [Author / George Sivulka]: The entire field of B2B AI companies for the next ten years will be built upon these differences:
[译文] [Author / George Sivulka]: 未来十年B2B AI公司的整个领域都将建立在这些差异之上:
📝 本节摘要:
本章介绍了“制度级智能”的第一个核心支柱:协同。作者通过一个思想实验指出,即使企业内部全是最优秀员工的克隆体,如果没有良好的管理、沟通与职责划分,组织也会陷入混乱。目前个人AI工具的普及正是如此,员工各自为战、输出互不相通,导致组织层面的工作流断裂。因此,未来的制度级智能必须建立一个强大的协调层,这甚至将催生出全新的“智能体管理(Agentic Management)”行业,来规范人类与智能体之间的分工、协作与价值衡量。
[原文] [Author / George Sivulka]: The Seven Pillars of Institutional Intelligence
[译文] [Author / George Sivulka]: 制度级智能(Institutional Intelligence)的七大支柱
[原文] [Author / George Sivulka]: 1. Coordination
[译文] [Author / George Sivulka]: 1. 协同(Coordination)
[原文] [Author / George Sivulka]: Individual AI creates chaos.
[译文] [Author / George Sivulka]: 个人AI(Individual AI)制造混乱。
[原文] [Author / George Sivulka]: Institutional AI creates coordination.
[译文] [Author / George Sivulka]: 制度级AI(Institutional AI)创造协同。
[原文] [Author / George Sivulka]: Let's begin with a thought experiment. Imagine you doubled your organization's headcount tomorrow with clones of only your best employees.
[译文] [Author / George Sivulka]: 让我们从一个思想实验开始。想象一下,如果明天你用你最优秀员工的克隆体将你组织的员工总数翻倍。
[原文] [Author / George Sivulka]: Each of these employees have minor differences, predilections, quirks, and perspectives (especially true if they're your best employees). If they're not sufficiently managed, if they're not sufficiently communicating, if their swim lanes, OKRs, roles and responsibilities are not well defined ... you've created chaos.
[译文] [Author / George Sivulka]: 这些员工中的每一个都有细微的差异、偏好、怪癖和观点(如果他们是你最优秀的员工,这点尤为真实)。如果他们没有得到充分的管理,如果他们没有充分地沟通,如果他们的业务范围(swim lanes)、目标与关键结果(OKRs)、角色和职责没有被很好地定义……你就会制造出混乱。
[原文] [Author / George Sivulka]: The organization, while measured on an individual basis, may be more productive, but thousands of agents (or humans) rowing in opposing directions creates a standstill at best, and destroys organizational harmony at worst.
[译文] [Author / George Sivulka]: 这个组织虽然以个人为基础衡量可能更具生产力,但成千上万个向着相反方向划船的智能体(agents)或人类,往好了说是造成停滞,往坏了说是破坏了组织的和谐。
[原文] [Author / George Sivulka]: This isn't hypothetical. It's happening right now in every organization that's adopted AI without a coordination layer. Every employee has their own ChatGPT habits, their own prompting styles, their own outputs that don't talk to anyone else's outputs. An org chart might exist, but the actual flow of AI-generated work says something else entirely.
[译文] [Author / George Sivulka]: 这不是假设。它现在正发生在一个没有协调层(coordination layer)却采用了AI的每个组织中。每个员工都有他们自己的ChatGPT习惯,他们自己的提示词风格(prompting styles),以及他们自己互不相通的输出结果。组织结构图(org chart)可能存在,但AI生成工作的实际流程说明的却是完全另一回事。
[原文] [Author / George Sivulka]: Fig. 2: Productive individuals (or agents) row in different directions alone. If left uncoordinated, chaos ensues.
[译文] [Author / George Sivulka]: 图2:高生产力的个体(或智能体)各自向不同的方向划船。如果任由其缺乏协调,混乱就会接踵而至。
[原文] [Author / George Sivulka]: Coordination is an absolute imperative, for humans and agents alike.
[译文] [Author / George Sivulka]: 协同是一个绝对的必要条件,对人类和智能体皆是如此。
[原文] [Author / George Sivulka]: Institutional intelligence will evolve into an entire “Agentic Management” industry focusing on agent roles and responsibilities, agent-to-agent and agent-to-human communication, and measuring agentic value (consumption based pricing alone doesn't cut it).
[译文] [Author / George Sivulka]: 制度级智能将演变为一个完整的“智能体管理(Agentic Management)”行业,专注于智能体的角色和职责、智能体与智能体以及智能体与人类之间的沟通,并衡量智能体价值(单靠基于消耗的定价模式是行不通的)。
📝 本节摘要:
本章指出了AI内容生成泛滥带来的新危机:个人AI极大地降低了生成内容的门槛,导致海量“AI垃圾(slop)”和噪音的产生。在信息爆炸的环境下,生成内容不再是难题,真正的挑战在于如何从指数级增长的噪音中筛选出真正有价值的“信号”。作者强调,未来的制度级AI必须依靠具备明确步骤和检查点的“确定性智能体(deterministic agents)”,而非不可预测的个人AI工具,来为组织进行信息过滤并创造实际的经济回报。
[原文] [Author / George Sivulka]: 2. Signal
[译文] [Author / George Sivulka]: 2. 信号(Signal)
[原文] [Author / George Sivulka]: Individual AI creates noise.
[译文] [Author / George Sivulka]: 个人AI制造噪音。
[原文] [Author / George Sivulka]: Institutional AI finds signal.
[译文] [Author / George Sivulka]: 制度级AI寻找信号。
[原文] [Author / George Sivulka]: Humans today are able to create, or rather generate, anything they can imagine: AI-essays, presentations, spreadsheets, photos, videos, songs, websites, and software. What a gift.
[译文] [Author / George Sivulka]: 如今的人类能够创造,或者更准确地说是生成,任何他们能想象到的东西:AI文章、演示文稿、电子表格、照片、视频、歌曲、网站和软件。这真是一份厚礼。
[原文] [Author / George Sivulka]: The issue is that almost everything generated by AI is complete slop. The proliferation of this AI slop has become so bad that some organizations are over-rotating and banning AI outputs altogether. This resonates personally... I run an AI company but ask our executive team not to use AI for any final written product. I can't stand the slop.
[译文] [Author / George Sivulka]: 问题在于,几乎所有由AI生成的东西都是彻头彻尾的垃圾(slop)。这种AI垃圾的激增已经变得如此糟糕,以至于一些组织正在矫枉过正,完全禁止使用AI输出。这让我深有同感……我经营着一家AI公司,但我要求我们的高管团队在任何最终的书面产品中不要使用AI。我无法忍受这些垃圾。
[原文] [Author / George Sivulka]: Imagine what the world of PE is quickly becoming. Last year, 10 deals may have come across your desk. This year, you'll receive 50 opportunities next quarter, each one AI-polished to perfection, and you have the same number of hours to find the one real deal.
[译文] [Author / George Sivulka]: 想象一下私募股权(PE)世界正在迅速变成什么样。去年,可能有10笔交易摆在你的办公桌上。今年,你将在下个季度收到50个机会,每一个都被AI打磨得完美无瑕,而你却只有同样多的时间来寻找那个唯一真正的交易。
[原文] [Author / George Sivulka]: Generating anything is no longer the problem. The problem, for any serious organization today, is generating and selecting the right thing. Finding the one good artifact, the one good deal, the signal in the noise, matters more and more in an AI-driven world. The key economic driver for the next decade will be uncovering the signal in the mountain of exponentially increasing slop.
[译文] [Author / George Sivulka]: 生成“任何东西”已不再是问题。今天,对于任何严肃的组织来说,问题在于生成并选择“正确的东西”。在一个由AI驱动的世界里,找到那一件优秀的作品、那一笔好交易、噪音中的信号,变得越来越重要。未来十年的关键经济驱动力将是从呈指数级增长的垃圾大山中发掘出信号。
[原文] [Author / George Sivulka]: Fig. 3: AI slop from individual productivity tools is proliferating at an exponentially increasing rate. Humans alone can't sort through the noise, and an institutional class of new AI products is needed
[译文] [Author / George Sivulka]: 图3:来自个人生产力工具的AI垃圾正以指数级增长的速度激增。单靠人类无法在噪音中进行梳理,需要一种制度级的新型AI产品。
[原文] [Author / George Sivulka]: Institutional-grade intelligence must find the signal, it must structure the noise to cut through slop, and it must be defined, deterministic, and auditable in the work it does.
[译文] [Author / George Sivulka]: 制度级别的智能必须找到信号,它必须对噪音进行结构化处理以穿透这些垃圾,而且它在执行工作时必须是明确的、确定性的和可审计的(defined, deterministic, and auditable)。
[原文] [Author / George Sivulka]: Whereas individual AI might emphasize the “always on” productivity of a Clawdbot exploring unpredictable ways to tend to one's 24/7 needs, i.e. a nondeterministic agent, institutional AI will rely upon the load-bearing predictability of deterministic agents. Agents that have predictable checkpoints, steps, and processes that they run will scale, will uncover signal, and through that signal drive returns via revenue for an organization.
[译文] [Author / George Sivulka]: 个人AI可能会强调一个像Claude机器人(Clawdbot)那样“始终在线”的生产力,探索不可预测的方式来照顾人们全天候的需求,即一种非确定性智能体(nondeterministic agent),而制度级AI将依赖于确定性智能体(deterministic agents)那具有承载力的可预测性。那些拥有可预测的检查点、步骤和运行流程的智能体将会实现规模化,将会发掘出信号,并通过该信号为组织带来以收入为形式的回报。
[原文] [Author / George Sivulka]: Fig 4. Matrix is a tool that uses the power of generative technology to cut through the noise. And in doing so, opens up a world of determined agents, with checkpoints.
[译文] [Author / George Sivulka]: 图4:Matrix是一个利用生成式技术的力量来穿透噪音的工具。通过这样做,它打开了一个拥有检查点的、确定性智能体的世界。
📝 本节摘要:
本章探讨了AI发展中的“偏见”新问题。作者指出,虽然过去的AI曾因社会政治偏见受到指责,但如今的基础模型为了迎合人类,已经被过度对齐(over-aligned),变成了一味讨好用户的“马屁精”。这种“个人AI”会无限放大个人偏见,甚至助长表现不佳员工的盲目自信,从而对组织的和谐与决策产生破坏性影响。相反,“制度级AI”应当扮演“唱反调者”的角色,追求客观事实,敢于提出质疑和挑战,并通过设立AI董事、AI审计员等制度化约束来过滤风险并维护组织标准。
[原文] [Author / George Sivulka]: 3. Bias
[译文] [Author / George Sivulka]: 3. 偏见(Bias)
[原文] [Author / George Sivulka]: Individual AI feeds bias.
[译文] [Author / George Sivulka]: 个人AI助长偏见。
[原文] [Author / George Sivulka]: Institutional AI creates objectivity.
[译文] [Author / George Sivulka]: 制度级AI创造客观性。
[原文] [Author / George Sivulka]: Concern around sociopolitical bias dominated AI discourse for years.
[译文] [Author / George Sivulka]: 多年来,围绕社会政治偏见的担忧主导了AI领域的话语。
[原文] [Author / George Sivulka]: The foundation model labs eventually circumvented the issue with enough RLHF to effectively turn all models into sycophants.
[译文] [Author / George Sivulka]: 基础模型实验室最终通过足够多的人类反馈强化学习(RLHF),有效地将所有模型变成了马屁精(sycophants),从而规避了这个问题。
[原文] [Author / George Sivulka]: Today, ChatGPT, Claude, etc. are so (overly) aligned that they'll agree with you on any topics within the Overton window (and sometimes slightly beyond, looking at you @Grok).
[译文] [Author / George Sivulka]: 如今,ChatGPT、Claude等模型被如此(过度)对齐(aligned),以至于它们会在奥弗顿之窗(Overton window)内的任何话题上赞同你(有时甚至稍微超出这个范围,说的就是你,@Grok)。
[原文] [Author / George Sivulka]: The discourse on sociopolitical bias has died down. A new problem has taken its place.
[译文] [Author / George Sivulka]: 关于社会政治偏见的讨论已经平息。一个新问题取而代之。
[原文] [Author / George Sivulka]: But this level of agreement—of over-alignment—on everything has become comically bad.
[译文] [Author / George Sivulka]: 但这种在所有事情上的高度一致——过度对齐——已经变得糟糕到了滑稽的地步。
[原文] [Author / George Sivulka]: It's become a meme in its own right ... Claude's reflexive “you're absolutely right!” regardless of whether or not you are, in fact, absolutely right.
[译文] [Author / George Sivulka]: 它本身已经成为了一个梗(meme)……无论你事实上是否真的绝对正确,Claude都会条件反射般地说“你绝对正确!”。
[原文] [Author / George Sivulka]: This sounds harmless. It is not.
[译文] [Author / George Sivulka]: 这听起来似乎无害。其实不然。
[原文] [Author / George Sivulka]: The loudest AI advocates inside many organizations may soon be the historically worst-performing employees. Think about why.
[译文] [Author / George Sivulka]: 许多组织内部声音最响亮的AI倡导者,可能很快就会是历史上表现最差的员工。想想为什么。
[原文] [Author / George Sivulka]: Organizations' worst employees, who receive little to no positive reinforcement every day, will soon have ASI agreeing with them.
[译文] [Author / George Sivulka]: 组织里那些每天几乎得不到任何正向反馈的最差员工,很快就会拥有同意他们观点的人工超级智能(ASI)。
[原文] [Author / George Sivulka]: They will whisper to themselves, “the smartest intelligence that has ever existed agrees with me. My manager is wrong.”
[译文] [Author / George Sivulka]: 他们会在心里对自己嘀咕:“有史以来最聪明的智能也同意我的看法。我的经理错了。”
[原文] [Author / George Sivulka]: This is intoxicating. It's also organizationally toxic.
[译文] [Author / George Sivulka]: 这是令人陶醉的。但它对组织也是有毒的。
[原文] [Author / George Sivulka]: Fig 5. Individual AI echo chambers fuel division, drawing two humans apart, a dynamic which at scale creating factions in an otherwise coherent organization.
[译文] [Author / George Sivulka]: 图5:个人AI的信息茧房(echo chambers)加剧了分歧,将两个人拉开距离,这种动态在规模化后会在一个原本紧密的组织中制造出不同的派系。
[原文] [Author / George Sivulka]: This highlights something important. These individual productivity tools reinforce the user.
[译文] [Author / George Sivulka]: 这突显了一件重要的事情。这些个人生产力工具在不断强化用户(reinforce the user)。
[原文] [Author / George Sivulka]: In reality the most important thing to reinforce is the truth.
[译文] [Author / George Sivulka]: 而在现实中,最重要被强化的应该是真相。
[原文] [Author / George Sivulka]: Organizations have evolved over thousands of years to build systems that counteract exactly this problem:
[译文] [Author / George Sivulka]: 组织经过数千年的演变,建立了正是为了抵消这个问题的各种系统:
[原文] [Author / George Sivulka]: * Investment committee meetings
[译文] [Author / George Sivulka]: * 投资委员会会议
[原文] [Author / George Sivulka]: * Third-party diligence
[译文] [Author / George Sivulka]: * 第三方尽职调查
[原文] [Author / George Sivulka]: * Boards of Directors
[译文] [Author / George Sivulka]: * 董事会
[原文] [Author / George Sivulka]: * The executive, legislative, and judicial branches of the US government
[译文] [Author / George Sivulka]: * 美国政府的行政、立法和司法分支
[原文] [Author / George Sivulka]: * Representative democracy, and democracy as a whole
[译文] [Author / George Sivulka]: * 代议制民主,以及整体的民主制度
[原文] [Author / George Sivulka]: Fig 6. Objectivity even attenuates the coordination problem, taking small differences and dampening vs. amplifying them.
[译文] [Author / George Sivulka]: 图6:客观性甚至能缓解协同问题,它吸收微小的差异并将其抑制,而不是将其放大。
[原文] [Author / George Sivulka]: Organizations rarely fail because people lack confidence. They fail because no one is willing, or able, to say no.
[译文] [Author / George Sivulka]: 组织很少因为人们缺乏自信而失败。它们失败是因为没有人愿意,或者能够说“不”。
[原文] [Author / George Sivulka]: Institutional AI must play that role.
[译文] [Author / George Sivulka]: 制度级AI必须扮演这个角色。
[原文] [Author / George Sivulka]: It will not be RLHF'ed into flattering users or echoing their beliefs, but to challenge their bias.
[译文] [Author / George Sivulka]: 它不会被人类反馈强化学习(RLHF)训练成去奉承用户或附和他们的信念,而是去挑战他们的偏见。
[原文] [Author / George Sivulka]: It will reinforce behavior when productive, and draw a hard line in realigning non-productive tendencies.
[译文] [Author / George Sivulka]: 它会在行为具有生产力时予以强化,并在纠正非生产力倾向时划定一条强硬的底线。
[原文] [Author / George Sivulka]: Thus, the most important agents inside organizations will not be “yes-men” but disciplined “no-men” that interrogate reasoning, surface risks, and enforce standards.
[译文] [Author / George Sivulka]: 因此,组织内部最重要的智能体将不会是“好好先生(yes-men)”,而是纪律严明的“反对者(no-men)”,它们会审问推理过程,揭示潜在风险,并执行标准。
[原文] [Author / George Sivulka]: Some of the most consequential future applications of AI will be built around institutional constraints: AI board members, AI auditors, AI third-party testing, AI compliance, and many more…
[译文] [Author / George Sivulka]: 未来AI一些最重要且影响深远的应用将围绕制度约束来构建:AI董事会成员、AI审计员、AI第三方测试、AI合规等等……
📝 本节摘要:
本章探讨了“制度级智能”的第四个核心支柱:边缘优势(Edge)。作者指出,个人AI侧重于追求广泛的通用性与基础使用率,而制度级AI则专注于在特定细分领域建立极致的竞争壁垒。尽管通用基础模型的能力正在飞速迭代,但针对特定目标构建的专业应用(如Midjourney、Elevenlabs)始终能在其垂直赛道保持领先。在金融等追求直接经济回报的领域,哪怕是1%的微小优势也能撬动数十亿美元的价值。因此,未来企业级AI的发展不是让基础模型替代一切,而是将通用大模型与高度专用的垂直智能体相结合。哪怕是未来的通用人工智能(AGI),也会需要专用工具来维持其在特定领域的极致优势。
[原文] [Author / George Sivulka]: 4. Edge
[译文] [Author / George Sivulka]: 4. 优势(Edge)
[原文] [Author / George Sivulka]: Individual AI optimizes for usage.
[译文] [Author / George Sivulka]: 个人AI为使用率(usage)而优化。
[原文] [Author / George Sivulka]: Institutional AI optimizes for edge.
[译文] [Author / George Sivulka]: 制度级AI为优势(edge)而优化。
[原文] [Author / George Sivulka]: The goalposts in AI evolve on a weekly and sometimes daily cadence. Foundation model companies, competing for every person and every organization, are rapidly iterating on capabilities.
[译文] [Author / George Sivulka]: AI领域的目标柱(goalposts)以每周甚至每天的节奏在演变。为了争夺每一个个人和每一个组织,基础模型(Foundation model)公司正在快速迭代其能力。
[原文] [Author / George Sivulka]: But in the classic innovators' dilemma, depth beats breath for specific applications every time:
[译文] [Author / George Sivulka]: 但在经典的创新者窘境(innovators' dilemma)中,对于特定应用而言,深度(depth)每一次都能击败广度(breath):
[原文] [Author / George Sivulka]: * It's @Midjourney's job to be slightly ahead on designed imagery.
[译文] [Author / George Sivulka]: * @Midjourney 的工作是在设计图像方面保持轻微领先。
[原文] [Author / George Sivulka]: * It's @Elevenlabsio's job to be slightly ahead on voice models.
[译文] [Author / George Sivulka]: * @Elevenlabsio 的工作是在语音模型方面保持轻微领先。
[原文] [Author / George Sivulka]: * And it's @DecagonAI's job to be always ahead on full-stack customer service experience...
[译文] [Author / George Sivulka]: * 而 @DecagonAI 的工作是在全栈客户服务体验上始终保持领先……
[原文] [Author / George Sivulka]: And while the foundation models will get close, the true edge matters for experts in their field. Many of the best designers use @Midjourney, many of the best voice AI companies will use @Elevenlabsio, etc … because even as the foundation models improve, the unyielding focus purpose-built applications have on driving their specific edge defines the edge itself.
[译文] [Author / George Sivulka]: 虽然基础模型会不断逼近,但真正的优势对于其领域的专家来说至关重要。许多最顶尖的设计师使用 @Midjourney,许多最顶尖的语音AI公司将使用 @Elevenlabsio,等等……因为即使基础模型在不断改进,专用(purpose-built)应用在推动其特定优势方面所展现出的不屈不挠的专注,本身就定义了这种优势。
[原文] [Author / George Sivulka]: As long as purpose-built solutions evolve too, the capabilities that matter for economic outcomes, for businesses, will always be with purpose-built products.
[译文] [Author / George Sivulka]: 只要专用解决方案也在进化,那些对企业的经济成果至关重要的能力,将永远属于专用产品。
[原文] [Author / George Sivulka]: This plays out to a tee in finance – the hottest area for LLM development right now. As soon as a capability is wide spread, it definitionally isn't going to help you beat the market. But if frontier technology can yield an ephemeral 1 percent niche advantage? That 1 percent can be levered into billion dollar outcomes.
[译文] [Author / George Sivulka]: 这一点在金融领域表现得淋漓尽致——这正是目前大型语言模型(LLM)开发最热门的领域。一旦某项能力被广泛普及,从定义上讲它就无法帮助你击败市场。但如果前沿技术能带来哪怕是短暂的1%的生态位优势(niche advantage)呢?这1%可以被杠杆化为价值十亿美元的回报。
[原文] [Author / George Sivulka]: Fig 7. The edge for any sufficiently specific task is defined by the institutional solutions you build on top of frontier technology.
[译文] [Author / George Sivulka]: 图7:对于任何足够具体的任务,其优势是由你在前沿技术之上构建的制度级解决方案所定义的。
[原文] [Author / George Sivulka]: Our users have always exceeded the frontier. Context windows in LLMs have grown from 4K to 1M tokens in four years. Some of our users process 30B tokens in a single job. We have line of sight to 100B-token jobs this year. Every time foundation model capabilities improve, we've already pushed further.
[译文] [Author / George Sivulka]: 我们的用户总是不断超越前沿。在过去的四年里,LLM的上下文窗口(Context windows)已经从4K增长到了100万(1M)个token。我们的一些用户在单个任务中处理高达300亿(30B)个token。今年我们有望看到处理1000亿(100B)个token的任务。每次基础模型的能力提升时,我们都已经推进得更远了。
[原文] [Author / George Sivulka]: Fig 8. Context windows, like other capabilities, are a moving goal post game. The last 3 years of context window evolution from the frontier labs, and at Hebbia.
[译文] [Author / George Sivulka]: 图8:与其他能力一样,上下文窗口是一场不断移动目标柱(moving goal post)的游戏。过去3年里前沿实验室以及Hebbia在上下文窗口方面的演变。
[原文] [Author / George Sivulka]: Usage for broad populations is important and worthwhile as a goal in itself, especially in onboarding employees to AI. But the future will not be people using ChatGPT/Claude or a domain-specific solution. It will be ChatGPT/Claude and a domain-specific solution.
[译文] [Author / George Sivulka]: 针对广泛人群的使用率(Usage)本身作为一个目标是重要且有价值的,特别是在引导员工使用AI方面。但未来不会是人们使用ChatGPT/Claude“或”某个特定领域解决方案的二选一。未来将是ChatGPT/Claude“与”特定领域解决方案的结合。
[原文] [Author / George Sivulka]: Institutional intelligence must leverage domain-specific, perhaps even task specific, agents.
[译文] [Author / George Sivulka]: 制度级智能必须利用特定领域(domain-specific)甚至特定任务(task specific)的智能体。
[原文] [Author / George Sivulka]: We ask ourselves a question that sounds absurd but isn't:
[译文] [Author / George Sivulka]: 我们问了自己一个听起来荒谬但其实不然的问题:
[原文] [Author / George Sivulka]: “What are the agents an AGI would choose to use as a shortcut? Even superintelligence would want purpose-built tools for specific domains.”
[译文] [Author / George Sivulka]: “通用人工智能(AGI)会选择使用哪些智能体作为捷径?即便是超级智能(superintelligence)也会想要针对特定领域的专用工具。”
[原文] [Author / George Sivulka]: The goalposts will always change in AI, and the organizations that leverage the true edge of capability are the organizations that will win. Everyone else is paying for a very expensive commodity.
[译文] [Author / George Sivulka]: AI领域的目标柱将永远在变,而那些能够利用真正前沿能力优势的组织将会获胜。其他所有人都在为一种极其昂贵的商品(commodity)买单。
📝 本节摘要:
本章阐述了“制度级智能”的第五个核心支柱:结果与收益导向(Outcomes)。当前市场上的大多数个人AI产品主要承诺节省时间和削减成本,但企业的核心诉求往往是实现收入增长。作者以软件开发和并购(M&A)为例,指出个人AI只是提升了单点效率(如更快地建模型或写代码),而制度级AI则致力于出售“业务转型”并直接创造收入(如精准定位交易对手)。因此,未来的核心价值将向“解决方案层(solution layer)”聚集,真正将技术与最终商业结果结合的制度级AI才能捕获最大的上升空间。
[原文] [Author / George Sivulka]: 5. Outcomes
[译文] [Author / George Sivulka]: 5. 结果(Outcomes)。
[原文] [Author / George Sivulka]: Individual AI saves time.
[译文] [Author / George Sivulka]: 个人AI节省时间。
[原文] [Author / George Sivulka]: Institutional AI scales revenue.
[译文] [Author / George Sivulka]: 制度级AI扩大收入。
[原文] [Author / George Sivulka]: @MaVolpi once told me something that reframed how I think about selling AI to the enterprise: “If you ask any CEO whether their first priority is cutting costs or scaling revenue, almost all would say revenue.”
[译文] [Author / George Sivulka]: @MaVolpi 曾对我说过一番话,重塑了我对向企业销售AI的看法:“如果你问任何一位CEO他们的首要任务是削减成本还是扩大收入,几乎所有人都会说是收入。”
[原文] [Author / George Sivulka]: Yet almost every AI product on the market today delivers cost-cutting, promising us to save time, do more with less, or replace headcount.
[译文] [Author / George Sivulka]: 然而,今天市场上几乎每一款AI产品提供的都是削减成本,向我们承诺节省时间、事半功倍,或取代员工编制(headcount)。
[原文] [Author / George Sivulka]: Institutional AI must deliver upside. And upside is a lot harder to commoditize than saved time.
[译文] [Author / George Sivulka]: 制度级AI必须带来上升空间(upside)。而上升空间要比节省的时间难商品化得多。
[原文] [Author / George Sivulka]: Take the example of agentic software development. Coding IDEs are some of the best individual AI productivity tools ever built, and they're already facing massive headwinds from Claude Code, another individual AI tool.
[译文] [Author / George Sivulka]: 以基于智能体(agentic)的软件开发为例。编程集成开发环境(IDEs)是有史以来构建的最好的个人AI生产力工具之一,但它们已经面临来自另一款个人AI工具 Claude Code 的巨大逆风。
[原文] [Author / George Sivulka]: Cognition is playing an entirely different game. Their most steadily growing business builds tech to sell transformations, not tools. I'd bet on that lasting power.
[译文] [Author / George Sivulka]: Cognition 正在玩一场完全不同的游戏。他们最稳定增长的业务是构建技术来销售转型(transformations),而不是工具。我押注于这种持久的生命力。
[原文] [Author / George Sivulka]: Pure software “is rapidly becoming uninvestable.” Pure services don't scale. The solution layer, marrying technology to outcomes, is where lasting value accumulates.
[译文] [Author / George Sivulka]: 纯软件“正迅速变得不再具备投资价值(uninvestable)”。纯服务无法规模化。将技术与结果(outcomes)结合起来的解决方案层(solution layer),才是持久价值累积的地方。
[原文] [Author / George Sivulka]: Or take M&A. Individual AI helps an analyst build a model faster. Institutional AI identifies the one counterparty worth pursuing out of a hundred, and expands that universe to a thousand. One saves time; the other generates revenue.
[译文] [Author / George Sivulka]: 或者以并购(M&A)为例。个人AI帮助分析师更快地建立模型。制度级AI则从一百个交易对手中识别出那唯一一个值得追求的,并将这个范围扩展到一千个。一个是在节省时间;另一个是在创造收入。
[原文] [Author / George Sivulka]: Fig 9. Foundation models companies are moving into the vertical app layer. Vertical app layer companies are moving to the solution layer.
[译文] [Author / George Sivulka]: 图9:基础模型公司正在进入垂直应用层(vertical app layer)。垂直应用层公司正在向解决方案层(solution layer)移动。
[原文] [Author / George Sivulka]: Moving “upstream” is the natural gravity of the market right now. Foundation models are moving to the app layer. App layer companies moving to the solution layer.
[译文] [Author / George Sivulka]: 向“上游”移动是目前市场自然的引力。基础模型正在向应用层移动。应用层公司正在向解决方案层移动。
[原文] [Author / George Sivulka]: Institutional intelligence is the solution layer. And the solution layer, where the outcomes live, will accumulate lasting value and capture the biggest upside.
[译文] [Author / George Sivulka]: 制度级智能就是解决方案层。而解决方案层,这个结果(outcomes)存在的地方,将累积持久的价值并捕获最大的上升空间。
📝 本节摘要:
本章探讨了“制度级智能”的第六个核心支柱:赋能与流程工程(Enablement)。作者指出,人类天生抗拒改变,即便面对明显的技术劣势,许多组织甚至高管依然会拒绝使用AI。因此,仅仅给员工提供一个AI工具(个人AI的做法)是远远不够的。未来的制度级AI需要真正懂得行业痛点的“流程工程”专家,将企业的具体业务流程编码进智能体中,并推动自上而下的变革管理。单纯懂技术的软件团队无法胜任这一点,只有真正理解业务领域的制度级解决方案,才能教导和赋能整个组织完成AI转型。
[原文] [Author / George Sivulka]: 6. Enablement
[译文] [Author / George Sivulka]: 6. 赋能(Enablement)
[原文] [Author / George Sivulka]: Individual AI gives you a tool.
[译文] [Author / George Sivulka]: 个人AI给你一个工具。
[原文] [Author / George Sivulka]: Institutional AI shows you how to use it.
[译文] [Author / George Sivulka]: 制度级AI教你如何使用它。
[原文] [Author / George Sivulka]: Humans, for all our ingenuity, are reluctant to change.
[译文] [Author / George Sivulka]: 尽管人类充满创造力,但我们仍不愿改变。
[原文] [Author / George Sivulka]: Believe it or not, there are still successful businesses in NYC that don't accept credit cards.
[译文] [Author / George Sivulka]: 信不信由你,在纽约市仍然有一些成功的企业不接受信用卡。
[原文] [Author / George Sivulka]: They're losing money, they know they're losing money, and they're still unflinching in that inertia.
[译文] [Author / George Sivulka]: 他们在亏钱,他们知道自己在亏钱,但他们依然在这种惯性中毫不动摇。
[原文] [Author / George Sivulka]: Similarly, for the indefinite future, employees somewhere, in some organizations, will refuse to use AI.
[译文] [Author / George Sivulka]: 同样地,在无限的未来里,某些地方、某些组织中的员工将会拒绝使用AI。
[原文] [Author / George Sivulka]: Making the transition from a human-only organization to an AI-first hybrid organization is going to be the lasting and defining challenge of the next decade.
[译文] [Author / George Sivulka]: 完成从纯人类组织向人工智能优先(AI-first)的混合型组织过渡,将是未来十年持久且具有决定性意义的挑战。
[原文] [Author / George Sivulka]: And in many cases, the most senior, and most important, levels of the organization will be the slowest to adopt.
[译文] [Author / George Sivulka]: 在许多情况下,组织中最高级、最重要的层级往往是采用新技术最慢的。
[原文] [Author / George Sivulka]: Fig 10. The highest levels of an organization– the furthest from “productivity tool activity” are often the slowest, and most important, players to adopting new technology.
[译文] [Author / George Sivulka]: 图10:组织的最高层——距离“生产力工具活动”最远的人——往往是采用新技术最慢,却也是最重要的参与者。
[原文] [Author / George Sivulka]: There is a reason that Palantir is the only “software” company that is still trading at extraordinary multiples amidst a trillion dollar selloff in technology stocks over the last two months.
[译文] [Author / George Sivulka]: 在过去两个月科技股遭抛售达万亿美元之际,Palantir 是唯一一家仍以极高倍数交易的“软件”公司,这是有原因的。
[原文] [Author / George Sivulka]: Palantir is one of the first true “process engineering” companies.
[译文] [Author / George Sivulka]: Palantir 是首批真正的“流程工程(process engineering)”公司之一。
[原文] [Author / George Sivulka]: Whether you call it “process engineering” or “writing Claude skills files,” institutional AI of the future will have an industry of encoding firm processes in agents and actualizing the change management required to put them in action.
[译文] [Author / George Sivulka]: 无论你称之为“流程工程”还是“编写 Claude 技能文件”,未来的制度级AI都将催生一个专注于将企业流程编码到智能体中,并实施所需的变革管理以将其付诸实践的行业。
[原文] [Author / George Sivulka]: Fig 11. There will be many chasms to cross for full organizational adoption. Each will have its challenges. Bringing processes online for AI across the entire organization will be a major driver.
[译文] [Author / George Sivulka]: 图11:要实现组织的全面采用,还有许多鸿沟需要跨越。每一个都会带来挑战。在整个组织内将流程为AI在线化将是一个主要的驱动力。
[原文] [Author / George Sivulka]: I'd warrant that process engineering will become arguably the most important “technology” in the near term.
[译文] [Author / George Sivulka]: 我敢保证,流程工程在短期内可能会成为最重要的“技术”。
[原文] [Author / George Sivulka]: And in process engineering, business and industry expertise—not software expertise—is most important.
[译文] [Author / George Sivulka]: 而在流程工程中,商业和行业专业知识——而非软件专业知识——才是最重要的。
[原文] [Author / George Sivulka]: Domain specific solutions beget expertise in the professionals doing the forward deployed engineering, the deployment, and the change management.
[译文] [Author / George Sivulka]: 特定领域的解决方案要求负责前置部署工程(forward deployed engineering)、实施以及变革管理的专业人员具备专业知识。
[原文] [Author / George Sivulka]: A top 3 bulge bracket bank that chose Hebbia for wall-to-wall deployment put it best: They were turned off from working with a big model lab, when they “had to explain what a CIM was to the team.”
[译文] [Author / George Sivulka]: 一家选择 Hebbia 进行全面(wall-to-wall)部署的排名前三的大型投行(bulge bracket bank)说得最好:当他们“不得不向团队解释什么是CIM(保密信息备忘录)时”,他们便打消了与大型模型实验室合作的念头。
[原文] [Author / George Sivulka]: Claude or GPT surely knew the domain, but the lab's team architecting the rollout did not...
[译文] [Author / George Sivulka]: Claude 或 GPT 肯定了解这个领域,但该实验室负责构建部署方案的团队却不了解……
[原文] [Author / George Sivulka]: That made all the difference.
[译文] [Author / George Sivulka]: 这就是天壤之别。
📝 本节摘要:
本章揭示了“制度级智能”的最后一个核心支柱:自发行动(Unprompted)。目前的个人AI高度依赖人类的“提示(prompt)”,但这恰恰将系统能力受限于组织中最薄弱的环节——人类本身,因为人类往往不知道该问什么,或何时该问。真正的制度级AI应该打破“一问一答”的被动模式,转为主动持续地监控全局数据,自发地发现那些无人察觉的风险与机遇。当AI不再需要人类下达指令时,全新的交互界面与工作方式将被彻底激发。
[原文] [Author / George Sivulka]: 7. Unprompted
[译文] [Author / George Sivulka]: 7. 自发行动(Unprompted)
[原文] [Author / George Sivulka]: Individual AI responds to human prompts.
[译文] [Author / George Sivulka]: 个人AI响应人类的提示(prompts)。
[原文] [Author / George Sivulka]: Institutional AI acts unprompted.
[译文] [Author / George Sivulka]: 制度级AI自发行动(unprompted)。
[原文] [Author / George Sivulka]: There's much chatter about agent-to-agent communications, and whether the businesses, software products, and institutions of the future even need humans at all.
[译文] [Author / George Sivulka]: 关于智能体与智能体之间的通信,以及未来的企业、软件产品和制度是否还需要人类,有着大量的讨论。
[原文] [Author / George Sivulka]: The better question, however, is whether AI agents of the future will need prompting at all.
[译文] [Author / George Sivulka]: 然而,一个更好的问题是,未来的AI智能体是否还需要提示(prompting)。
[原文] [Author / George Sivulka]: Prompting an AGI is like hooking an electric motor into a power loom. It's fundamentally, irrevocably constrained by the weakest link in the organizational supply chain—us.
[译文] [Author / George Sivulka]: 给通用人工智能(AGI)提供提示,就像把电动机连接到动力织布机上一样。它从根本上、不可挽回地受制于组织供应链中最薄弱的环节——我们。
[原文] [Author / George Sivulka]: Humans hardly know the right questions to ask, let alone when to ask them.
[译文] [Author / George Sivulka]: 人类几乎不知道该问什么正确的问题,更不用说什么时候去问了。
[原文] [Author / George Sivulka]: The most valuable work AI can do is the work nobody thinks to ask for. AI should find the risk that nobody flagged, the counterparty nobody thought of, and the sales pipeline that nobody knew was there.
[译文] [Author / George Sivulka]: AI能做的最有价值的工作,是那些没有人想到去要求做的工作。AI应该发现没有人标记的风险,没有人想到的交易对手,以及没有人知道其存在的销售渠道。
[原文] [Author / George Sivulka]: This will blow open the manifold of AI use cases.
[译文] [Author / George Sivulka]: 这将彻底打开AI应用场景的广阔天地。
[原文] [Author / George Sivulka]: An unprompted system continuously watches incoming data across the entire portfolio. It detects that one company's working capital cycle has quietly deteriorated for three consecutive months, cross-references that against covenant thresholds in the credit agreement, and flags the operating partner before anyone at the fund has opened the PDF.
[译文] [Author / George Sivulka]: 一个无需提示的系统会持续监控整个投资组合中不断涌入的数据。它检测到某家公司的营运资金周期已连续三个月悄然恶化,将其与信贷协议中的契约阈值进行交叉比对,并在基金中任何人打开PDF之前向运营合伙人发出预警。
[原文] [Author / George Sivulka]: When you remove the need for humans to prompt AI, new interfaces and new ways of working emerge. We @Hebbia have some strong opinions here. To be continued.
[译文] [Author / George Sivulka]: 当你消除人类向AI提供提示的需求时,新的交互界面和新的工作方式就会涌现。我们 @Hebbia 在这方面有一些强烈的观点。未完待续。
📝 本节摘要:
本章作为全文的总结,重申了个人AI与制度级AI并非非此即彼的对立关系,而是“强强联合”的必然趋势。个人AI将作为普及工具,带领大多数企业初步体验AI的魔力;而在此基础上,企业迫切需要构建解决特定领域问题的制度级AI。作者最后呼应了文章开篇1890年代纺织厂的隐喻:在这个已经拥有了“电力”(基础AI技术)的时代,真正的赢家将是那些率先“重塑工厂”(重构组织与工作流程)的人。
[原文] [Author / George Sivulka]: Conclusion
[译文] [Author / George Sivulka]: 结语
[原文] [Author / George Sivulka]: None of this negates the need for chatbots, agents, and individual AI as a whole.
[译文] [Author / George Sivulka]: 所有这些并没有否定对聊天机器人、智能体以及整体个人AI的需求。
[原文] [Author / George Sivulka]: Individual AI will be the vector by which the majority of the world's businesses first experience the transformative magic of AI.
[译文] [Author / George Sivulka]: 个人AI将成为世界上大多数企业首次体验AI变革魔力的载体。
[原文] [Author / George Sivulka]: Driving for usage, and generalizable ease of use, is the key first step to the change management to build an AI first economy.
[译文] [Author / George Sivulka]: 推动使用率和具备普适性的易用性,是构建“AI优先(AI first)”经济所需的变革管理的关键第一步。
[原文] [Author / George Sivulka]: But there is an obvious, urgent, and gaping need for institutional intelligence at the same time.
[译文] [Author / George Sivulka]: 但与此同时,对制度级智能的需求也是明显、迫切且存在巨大缺口的。
[原文] [Author / George Sivulka]: Every organization in the future will have a chatbot from a big lab.
[译文] [Author / George Sivulka]: 未来的每个组织都将拥有一个来自大型实验室的聊天机器人。
[原文] [Author / George Sivulka]: And every organization will have institutional AI purpose-built for domain-specific problems—institutional AI that individual AI will leverage as the key tool in its own tool belt.
[译文] [Author / George Sivulka]: 而且每个组织也都将拥有专为特定领域问题构建的制度级AI——而个人AI会将其作为自身工具箱里的核心工具来加以利用。
[原文] [Author / George Sivulka]: The “better together” story for institutional AI and individual AI is inevitable.
[译文] [Author / George Sivulka]: 制度级AI与个人AI“强强联合(better together)”的故事是必然的。
[原文] [Author / George Sivulka]: But remember the lesson of the 1890s textile mills.
[译文] [Author / George Sivulka]: 但请记住1890年代纺织厂的教训。
[原文] [Author / George Sivulka]: The factories that electrified first lost to those who redesigned the floor.
[译文] [Author / George Sivulka]: 那些最先实现电气化的工厂,输给了那些重新设计了车间的工厂。
[原文] [Author / George Sivulka]: We have our electricity. It's time to redesign our factories.
[译文] [Author / George Sivulka]: 我们已经拥有了电力。是时候重新设计我们的工厂了。
[原文] [Author / George Sivulka]: Thanks to @aleximm and @WillManidis for proofreading, and to Will for his “ Tool Shaped Objects” essay which helped inspire this piece.
[译文] [Author / George Sivulka]: 感谢 @aleximm 和 @WillManidis 的校对,也感谢 Will 的《工具形状的物体(Tool Shaped Objects)》一文为本文提供了灵感。
(注:
📝 本节摘要:
本章收录了文章发布后读者的核心评论与互动。读者 Bob Pulver 指出,脱离了组织协同的个人生产力仅仅是一种“虚荣指标”,他强调了在从“个人”向“制度”过渡时,必须重新衡量AI成熟度,并在设计上坚持以人为本。读者 R B 则产生共鸣,认为企业不仅需要AI战略,更需要一个能驱动和维持AI一致采用的基础设施(作为变革管理的一部分),并高度赞同了文章对“偏见”问题的重新审视。
[原文] [System]: Discussion about this post
[译文] [System]: 关于本文的讨论
[原文] [System]: Comments Restacks
[译文] [System]: 评论 转帖(Restacks)
[原文] [Reader / Bob Pulver]: Bob Pulver
[译文] [Reader / Bob Pulver]: Bob Pulver
[原文] [System]: Mar 12
[译文] [System]: 3月12日
[原文] [System]: Liked by George Sivulka
[译文] [System]: George Sivulka 赞了
[原文] [Reader / Bob Pulver]: Such an incredibly important post, thank you for laying this out so clearly.
[译文] [Reader / Bob Pulver]: 这是一篇极其重要的文章,感谢你如此清晰地将其阐述出来。
[原文] [Reader / Bob Pulver]: I have been talking about this for the past 18 months on my podcast (and almost everywhere else).
[译文] [Reader / Bob Pulver]: 过去18个月里,我一直在我的播客(以及几乎所有其他地方)谈论这个话题。
[原文] [Reader / Bob Pulver]: Individual productivity became a new vanity metric, and told us nothing about how teams, departments, or organizations were doing.
[译文] [Reader / Bob Pulver]: 个人生产力成为了一种新的虚荣指标(vanity metric),它丝毫不能告诉我们团队、部门或组织的表现如何。
[原文] [Reader / Bob Pulver]: One (human) cog going 10x while others go 1-2x means something is going to break.
[译文] [Reader / Bob Pulver]: 一个(人类)齿轮以10倍速运转,而其他齿轮以1-2倍速运转,意味着有些东西将会崩溃。
[原文] [Reader / Bob Pulver]: You have to be measuring AI readiness and maturity with different attributes as you go from individual to institution.
[译文] [Reader / Bob Pulver]: 当你从个人走向制度时,你必须用不同的属性来衡量AI的准备度(readiness)和成熟度(maturity)。
[原文] [Reader / Bob Pulver]: Only two things I would add are 1) we need responsibility and human-centricity by design to mitigate risk and build things properly the first time, and 2) this also points to the importance of capitalizing on the institutional collective intelligence for making better decisions and architecting strategic work.
[译文] [Reader / Bob Pulver]: 我要补充的只有两点:1)我们需要在设计上注重责任和以人为本(human-centricity),以降低风险并一次性把事情做对;2)这也指出了利用制度级集体智能(institutional collective intelligence)来做出更好决策和构建战略性工作的重要性。
[原文] [System]: Reply
[译文] [System]: 回复
[原文] [System]: Share
[译文] [System]: 分享
[原文] [System]: 2 replies
[译文] [System]: 2条回复
[原文] [Reader / R B]: R B
[译文] [Reader / R B]: R B
[原文] [System]: Mar 12
[译文] [System]: 3月12日
[原文] [System]: Liked by George Sivulka
[译文] [System]: George Sivulka 赞了
[原文] [Reader / R B]: Thank you for this article.
[译文] [Reader / R B]: 感谢这篇文章。
[原文] [Reader / R B]: What stood out most for me and how I interpret this is that even more important than an AI strategy, companies need an infrastructure that drives, sustains and ensures consistent adoption of AI.
[译文] [Reader / R B]: 对我来说最突出的一点,以及我对此的理解是,比AI战略更重要的是,公司需要一个能够驱动、维持并确保AI得到一致采用的基础设施(infrastructure)。
[原文] [Reader / R B]: I view this as part of what normally would be a change management strategy.
[译文] [Reader / R B]: 我将其视为通常属于变革管理战略(change management strategy)的一部分。
[原文] [Reader / R B]: Also, thank you for including the critical component of bias!
[译文] [Reader / R B]: 另外,感谢你囊括了关于偏见(bias)这一关键组成部分!
[原文] [Reader / R B]: I feel like the topic used to be a central part of most conversations around AI but it has slowly disappeared as the race to proliferate has taken center stage.
[译文] [Reader / R B]: 我觉得这个话题曾经是大多数围绕AI对话的中心部分,但随着技术扩散的竞赛占据舞台中心,它已慢慢消失了。
[原文] [System]: Reply
[译文] [System]: 回复
[原文] [System]: Share
[译文] [System]: 分享
[原文] [System]: 20 more comments...
[译文] [System]: 还有20条评论...
(注:这部分为文章底部的读者互动区。通过引入真实的读者反馈,也完美呼应了前文中作者对于组织协作、偏见和变革管理的探讨。