當(dāng)模型能學(xué)會(huì)一家公司所有的專業(yè)判斷時(shí)，這家公司還剩什么？

微軟CEO薩蒂亞·納德拉近日在社交平臺(tái)上發(fā)了一篇文章，標(biāo)題是《A frontier without an ecosystem is not stable》。

馬斯克隨后回復(fù)一個(gè)詞耐人尋味：Interesting。

納德拉這篇文章的核心議題是，AI驅(qū)動(dòng)的經(jīng)濟(jì)中，企業(yè)會(huì)變成什么樣。他判斷，這一輪變化和以往的平臺(tái)遷移不同。過去用數(shù)字系統(tǒng)放大人力資本，這一次則可以在人和數(shù)字系統(tǒng)之間建立真正的認(rèn)知回路，這意味著模型可以持續(xù)吸收人和組織的專業(yè)知識(shí)，并將其商品化。

模型正在商品化組織能力

具體來說，上一輪數(shù)字化是“工具放大人”，軟件提升效率，但創(chuàng)造價(jià)值的核心仍然是人的經(jīng)驗(yàn)、判斷、關(guān)系和對(duì)行業(yè)的理解。原文的說法是：“In the past, we used digital systems to enhance human capital.”

這一輪AI不同之處在于，模型能讀文檔、總結(jié)流程、模仿判斷、參與決策，原本屬于企業(yè)內(nèi)部的專業(yè)知識(shí)第一次出現(xiàn)了被抽離、壓縮、標(biāo)準(zhǔn)化、再商品化的可能。原文的措辭是“AI models can continuously absorb the expertise of humans and organizations and commoditize it”，吸收并商品化的是專業(yè)知識(shí)和組織能力，不是“人”本身。

移動(dòng)互聯(lián)網(wǎng)時(shí)代也好，云計(jì)算時(shí)代也好，企業(yè)買到的都是工具、渠道和基礎(chǔ)設(shè)施，這些不會(huì)動(dòng)搖公司的核心能力。納德拉認(rèn)為AI讓這層邊界開始松動(dòng)，模型在長時(shí)間學(xué)習(xí)企業(yè)文檔、歷史項(xiàng)目、客戶交互、流程反饋和結(jié)果評(píng)估之后，可能從輔助工具變成可重復(fù)調(diào)用的組織能力容器。

原文的說法是，他擔(dān)心的不只是某個(gè)數(shù)字工具的使用方式，而是“how organizations continue to learn, build IP, differentiate, and thrive”。

如果企業(yè)把數(shù)據(jù)、任務(wù)和工作流持續(xù)喂給外部模型，短期效率提升，長期議價(jià)權(quán)下降。模型越強(qiáng)，企業(yè)越依賴；企業(yè)越依賴，模型方就越有能力吸納原本分散在各行業(yè)的利潤。

納德拉的原話是：“The last thing any of us want is a world where every company across every sector is ceding value to a few models that eat everything they see.”所謂“沒有生態(tài)系統(tǒng)的前沿是不穩(wěn)定的”，即只有模型層獨(dú)贏的AI產(chǎn)業(yè)結(jié)構(gòu)長期撐不住。

納德拉擔(dān)心所有企業(yè)把價(jià)值讓渡給少數(shù)模型。如果所有經(jīng)濟(jì)回報(bào)只歸少數(shù)模型，政治經(jīng)濟(jì)體系不會(huì)容忍，也沒有社會(huì)許可讓AI掏空整個(gè)行業(yè)。

原文用了全球化的類比，第一波全球化中，整個(gè)工業(yè)經(jīng)濟(jì)被外包掏空，GDP數(shù)字看起來沒問題，但實(shí)際 displacement 是真實(shí)的，后果至今仍在。納德拉說不要讓這種動(dòng)態(tài)進(jìn)入AI時(shí)代，“Let us not bring that dynamic into the AI era, with a small number of AI systems capturing all the economic returns, while entire industries find their knowledge commoditized right out from underneath them.”

企業(yè)該怎么辦

文章中引人注意的是“human capital”和“token capital”這對(duì)概念。

人力資本原文定義為“the knowledge, judgment, relationships, ingenuity, and pattern recognition of its people”，專業(yè)知識(shí)、判斷力、人際關(guān)系、獨(dú)創(chuàng)性和模式識(shí)別。Token資本是“the firm's AI capability it builds and owns”，企業(yè)構(gòu)建并擁有的AI能力。

納德拉強(qiáng)調(diào)人力資本不會(huì)因Token資本增長而貶值，反而更值錢，因?yàn)槿说哪軇?dòng)性是Token資本增長的驅(qū)動(dòng)力。原文原話：“Human capital does not become less valuable as token capital grows. It only becomes more valuable!”

人的角色是設(shè)定目標(biāo)、跨領(lǐng)域串聯(lián)、建立關(guān)系、識(shí)別最重要的模式，原文原文是“Humans will set ambitious goals, connect dots across domains, build relationships, and recognize patterns that matter most.”沒有人的方向，算力只是在空轉(zhuǎn)。

需要說明的是，文中提到的“納德拉前幾天剛批過公司內(nèi)部的tokenmaxxing現(xiàn)象”指的是納德拉在6月11日 Hard Fork播客上的另一段發(fā)言，不是這篇文章的內(nèi)容。他在那場(chǎng)播客中說不要拿前沿模型解決非前沿問題。Token資本不是算力競(jìng)賽，而是企業(yè)把流程、反饋、知識(shí)、上下文和任務(wù)拆解成能被模型學(xué)習(xí)、調(diào)用和迭代的組織智能的能力，這個(gè)定義來自本篇文章。

真正的機(jī)會(huì)不是選最好的模型，而是在模型之上構(gòu)建學(xué)習(xí)回路，讓人力資本和Token資本復(fù)利增長。原文原話：“The real opportunity is not in picking the best model but instead in building a learning loop on top of models where human capital and token capital compound.”你可以外包任務(wù)甚至崗位，但絕不能外包學(xué)習(xí)。原文原話：“You can offload a task, or even a job, but can never offload your learning.”

納德拉提出了三條技術(shù)路徑，私有評(píng)估集（private evals）檢驗(yàn)?zāi)Ｐ褪欠裨跇I(yè)務(wù)實(shí)際目標(biāo)上改進(jìn)；私有強(qiáng)化學(xué)習(xí)環(huán)境（private reinforcement learning environments）讓模型基于組織內(nèi)部的真實(shí)數(shù)據(jù)變強(qiáng)；知識(shí)庫（knowledge base）讓組織記憶可查詢、token使用更高效。

這個(gè)學(xué)習(xí)回路成為企業(yè)新的IP，原文稱之為“a hill climbing machine”，而且會(huì)復(fù)利增長，每一條改進(jìn)的工作流產(chǎn)生更好的訓(xùn)練信號(hào)，加速積累企業(yè)獨(dú)有的隱性知識(shí)。企業(yè)應(yīng)該能換掉“通用模型”而不丟失嵌入學(xué)習(xí)系統(tǒng)中的“老員工專業(yè)知識(shí)”，這是納德拉給出的企業(yè)控制力和主權(quán)的關(guān)鍵測(cè)試。

組織必須建立自己的學(xué)習(xí)回路

納德拉因此主張優(yōu)先建前沿生態(tài)而非前沿模型，讓價(jià)值在各公司、各行業(yè)、各國之間廣泛流動(dòng)。原文原話：“Our priority has to be building a frontier ecosystem, not just a frontier model, so value flows broadly across every company, every industry, and every country.”他還提到了自己的信條，平臺(tái)讓頂層創(chuàng)造的價(jià)值超過自身捕獲的價(jià)值，“This is the ethos I've grown up with where platforms enable more value on top than is captured inside.”

納德拉不是產(chǎn)業(yè)外的中立觀察者，微軟既做前沿模型，也有云基礎(chǔ)設(shè)施、開發(fā)工具、企業(yè)辦公入口、身份和安全體系。對(duì)微軟來說，理想的產(chǎn)業(yè)結(jié)構(gòu)不是“只有模型最值錢”，而是模型、云、工具、工作流、安全、數(shù)據(jù)和應(yīng)用組成多層生態(tài)。

這篇文章給出的答案有微軟立場(chǎng)，前沿模型必須有，但不能成為唯一的價(jià)值中心。企業(yè)必須擁抱AI，但不能只當(dāng)模型的使用者。組織必須建立自己的學(xué)習(xí)回路，把人的判斷和機(jī)器的能力綁定在一起。如果市場(chǎng)真的認(rèn)為未來所有經(jīng)濟(jì)剩余歸極少數(shù)模型提供者，那平臺(tái)層、軟件層、協(xié)作層、企業(yè)IT層的估值邏輯都要重寫。

納德拉強(qiáng)調(diào)生態(tài)，部分原因是在阻止“所有利潤歸模型層”的敘事變成共識(shí)。

馬斯克回復(fù)“Interesting”之所以引起注意，是因?yàn)樗瑯诱莆漳Ｐ汀①Y本和公眾影響力，卻只用一個(gè)詞回應(yīng)了“別讓少數(shù)AI系統(tǒng)拿走全部回報(bào)”的主張。

“A frontier without an ecosystem is not stable.”這句話能不能成為行業(yè)共識(shí)還不確定，但它指出了一個(gè)越來越難繞開的現(xiàn)實(shí)，AI不只是把軟件再升級(jí)一次，它正在重新決定企業(yè)把什么留在自己體內(nèi)、把什么交給模型層、誰能從這場(chǎng)轉(zhuǎn)移中持續(xù)獲利。

未來分出高下的，也許不是誰接入了最多的模型，而是誰能把人的判斷留住，同時(shí)把組織經(jīng)驗(yàn)轉(zhuǎn)化成自己擁有、自己治理、自己能復(fù)利的Token資本。納德拉原文最后一句話是：“And it is the stable equilibrium we should build together.”

以下為納德拉X原文英文版：

I've been thinking a lot about the future of the firm in an AI-driven economy.

This transition is different than any previous platform shift. In the past, we used digital systems to enhance human capital. This is the first time we can create a real cognitive loop between people and digital systems. That is a mind-bender, because it changes how we even conceptualize work inside an enterprise.

What is at stake is not some digital tool or system and its use, but how organizations continue to learn, build IP, differentiate, and thrive in a world where AI models can continuously absorb the expertise of humans and organizations and commoditize it.

Every company is going to have to build what I think of as human capital and token capital. Human capital comprises the knowledge, judgment, relationships, ingenuity, and pattern recognition of its people, while token capital is the firm’s AI capability it builds and owns.

Importantly, human capital does not become less valuable as token capital grows. It only becomes more valuable! I believe human agency will be the driver of token capital growth. Humans will set ambitious goals, connect dots across domains, build relationships, and recognize patterns that matter most. Without human direction, you have compute running in circles.

This means the real opportunity is not in picking the best model but instead in building a learning loop on top of models where human capital and token capital compound. You can offload a task, or even a job, but you can never offload your learning. The future of the firm is the ability to compound that learning across people and AI.

This requires a new architectural approach where every business is able to build agentic systems that improve over time, while still retaining control over their IP. A company should be able to switch out a “generalist” model without losing the “company veteran” expertise built into their learning system. This is the key “test” of your control and sovereignty in the era ahead.

Companies need to turn their workflows, domain knowledge, and accumulated judgment into AI systems that improve with each use. Private evals should capture whether a model is actually improving against outcomes that matter to the business (not just external benchmarks!). Private reinforcement learning environments should let models grow stronger on real traces from inside the organization. Its knowledge base makes institutional memory queryable and use of tokens more efficient.

This loop becomes the new IP of the firm. I think of it as a hill climbing machine. And unlike most assets, it compounds. Every improved workflow generates better training signal, which accelerates the accumulation of tacit knowledge unique to the firm. The companies that build this early will have an advantage that is hard to replicate, regardless of any new individual model capability.

The last thing any of us want is a world where every company across every sector is ceding value to a few models that eat everything they see. If all the value is accrued by only a few models, the political economy will simply not tolerate it. There is no societal permission for an AI future that hollows out entire industries.
Think about what happened in the first phase of globalization where entire industrial economies were hollowed out by outsourcing. The GDP numbers looked fine on the surface, but the displacement was real and the consequences are still being felt. Let us not bring that dynamic into the AI era, with a small number of AI systems capturing all the economic returns, while entire industries find their knowledge commoditized right out from underneath them.

In my view, our priority has to be building a frontier ecosystem, not just a frontier model, so value flows broadly across every company, every industry, and every country. One where every organization can own the learning loop that encodes its institutional knowledge, compounding its human and token capital.
This is the ethos I’ve grown up with where platforms enable more value on top than is captured inside, and where every company can continuously innovate and build value of its own.

When that happens, companies will create value for themselves and for the economy around them. Employees will see their expertise amplified and their judgment become part of systems that make it replicable and scalable and the benefits accrue to the companies and communities around them.
That is how companies drive value for themselves and the broader economy. And it is the stable equilibrium we should build together.

（本文首發(fā)鈦媒體APP，作者 | 硅谷Tech_news，編輯 | 焦燕）