We may be bidding farewell to the "attention economy," which centers around "traffic" as its core asset, and moving towards an "agency economy," where "trust" becomes the core asset.

Author: Zhang Peng

Recently, OpenAI released the ChatGPT Pulse feature, which has sparked considerable discussion in the community. Sam Altman's description is clear: an AI that can work for you at night, continuously thinking about your interests, related data, recent conversations, and more, preparing some things you might be interested in every morning. If you tell ChatGPT what is important to you, it can perform exceptionally well.

Among them, the word "more" used by Sam refers, according to the official information on OpenAI ChatGPT Pulse, to the ability of Pulse to access Gmail and Google Calendar with user authorization, "to provide more contextual information for more relevant suggestions. After connecting to the calendar, ChatGPT may draft example meeting agendas, remind you to buy birthday gifts, or recommend restaurants for your upcoming trips." It is foreseeable that in the future, Pulse will likely connect to more applications to obtain richer user context.

On the surface, this seems like a smarter good morning greeting. But if we only see it as a new feature, we may miss some new trends implied behind it.

The emergence of Pulse is, in my view, worth paying attention to. This is OpenAI's attempt to define a brand new, deeper human-machine relationship. It first challenges the human-computer interaction paradigm that we have become accustomed to over the past decade. Behind this also brings a new possibility: the business empire model centered around "traffic" is finally beginning to loosen.

01 From Passive to Active

To understand Pulse, we must step out of the "function" box and see the changes at the underlying paradigm level.

First, it is an effort to shift from the role of a "passive tool" to that of an "active partner." "Proactivity" as a characteristic of this generation of AI technology is already a consensus, but OpenAI has successfully implemented it in Pulse. In the past, our interaction with ChatGPT was essentially a "question-and-answer" model. We were the active party, and the AI was the passive one. It was like a powerful magic box, but you needed to clearly know what you wanted and actively invoke it. Pulse attempts to reverse this relationship, making AI a partner that can "think proactively" and "serve actively." This shift from passive to active is a key leap in the evolution of human-computer interaction.

We can examine this new entity, Pulse, using the "AI Native Three Questions" framework from my speech at this year's FounderPark AGI Playground conference:

What is the new goal of Pulse? The goal of Pulse is no longer to "accurately answer a user's single question," but to "continuously optimize the user's overall experience and efficiency." This is an elevation from solving "point problems" to optimizing "overall states."

What is the new pipeline of Pulse? Its data processing flow has fundamentally changed. It is no longer a simple "question -> process -> answer." Instead, it is a continuous, background-running loop: "multi-dimensional data input -> autonomous integration and thinking -> proactive generation of value packages -> user feedback -> reinforcement learning." This constitutes a "data flywheel" that can continuously and proactively "push" the product side to turn.

What is the new value model of Pulse? Its value is no longer measured solely by the quality of single answers, but is reflected in the depth of data you are willing to entrust to it and the strength of the relationship established. The more you authorize, the deeper the investment in this "relationship," and the more unique value it creates for you, the higher your switching costs.

These changes also signify a new evolution of information service models. If Google's search represents the 1.0 era of "people finding information," and ByteDance's recommendation engine represents the 2.0 era of "information finding people," then what Pulse represents may be the beginning of a 3.0 era of "value generated for me" (Value Generation). It does not provide ready-made information cans but generates new value in real-time based on your unique context.

02 The Possibility of Transitioning from "Attention Economy" to "Agency Economy"

As the relationship between humans and AI is reshaped, the values and business models it can carry will also be reconstructed. Pulse paints a prototype of a "personal operating system" centered around "you" and a new possibility for economic models.

First, the innovation of agency: "Agent of Platform" vs. "Agent of You."

Today, all recommendation algorithms that claim to "understand you" are essentially "agents of the platform" (Agent of Platform). No matter how personalized they are, their fundamental KPIs are platform interests—retention, duration, conversion rates, GMV… The algorithm acts like a "top salesperson" placed by the platform beside you; its "goodness" towards you is to make the platform "better."

The emergence of products like Pulse signifies a revolution of "stance." The ultimate form of AI Agent should be a "your agent" (Agent of You) that serves you completely. Its only KPI should be your success and satisfaction. Every calculation and push it makes must withstand the ultimate question: "Does this align with your best interests?"

This role switch from "working for the platform" to "working for you" may seem small, but it shifts the entire pivot of business logic.

Second, the upgrade of the value chain: from "information distribution" to "service orchestration."

The role switch inevitably brings a qualitative change in the way value is delivered.

In the past, platforms built data silos on your phone, pushing information to you only within their business (e-commerce, content, social). A true "personal operating system" will break down these walls and become the "central dispatcher" for all your digital services.

A true "Personal OS" has "networked" capabilities. It will cross the boundaries of apps, becoming the "central dispatcher" for all your digital services, delivering not "information," but "services." It won't just recommend a travel guide; it will formulate and execute a travel plan for you: checking your calendar app for availability, connecting to travel apps to search and book flights, opening map apps to plan routes, and even integrating with budgeting apps to analyze and manage expenses.

In this "you-centered personal operating system," each isolated app on your phone is activated into "capability blocks" that can be called upon by the AI Agent at any time. It compresses complex task flows that previously required users to manually switch between multiple applications, copy, and paste into a single natural language conversation. Its theoretical upper limit equals the sum of all digital services. It accomplishes not "information distribution," but "service orchestration," thereby expanding the "new bandwidth" for creating value for users.

Third, the evolution of business models: from "attention economy" to "agency economy."

When the AI's stance aligns with the user, and the way value is delivered becomes service, the disruption of business models becomes inevitable. We may be bidding farewell to the "attention economy," centered around "traffic" as its core asset, and moving towards an "agency economy," where "trust" becomes the core asset.

The core of the old model is "traffic monetization," where platforms attract your attention through free content and then package it to sell to advertisers. The fundamental contradiction of this model lies in the frequent conflict between advertisers' needs and users' needs. In the "agency economy," the business model will become purer and more consistent: subscription model: you pay to hire this AI to work for you, which is the most direct agency relationship. It could even be result-oriented value sharing: when the AI creates quantifiable value for you through "service orchestration" (for example, helping you snag cheaper flights or achieving excess returns on your investment portfolio), it and the string of agents it "orchestrated" receive a pre-agreed percentage of commission or share, which is then internally distributed. Ultimately, the AI must continuously prove that it is serving your best interests for you to be willing to keep paying.

In this model, "delivering results" replaces "free use," and "trust value" replaces "traffic value," which will change the nature of the relationship between products and users and their user value.

03 Are Only Giants the Winners?

So an interesting question is: why haven't giants like Apple and WeChat, which have vast user data, rushed to create Pulse? Will they eventually create Pulse?

I have discussed this topic with some friends inside these two companies, and the answer is: "very cautious." This caution is not only about the compliance risks of user data but also reflects a profound challenge—I call it the challenge of establishing a "trust-fulfillment flywheel."

User-authorized data is the input of "trust"; AI delivering unexpected value is the output of "fulfillment." This is a crucial yet extremely fragile positive cycle.

This precisely explains the giants' dilemma. First, their size is too large; any "fulfillment" failure could be massive. The "trial-error-correction" costs of large products in big companies are too high compared to startups. Therefore, they tend to be cautious and adopt the so-called "better to be the last in the world."

A deeper reason is that using the "user trust" gained in business A to "fulfill" business B can easily trigger user vigilance and aversion towards the giants, even leading to antitrust issues, even if the starting point is good.

As a new challenger, OpenAI can be understood as using a non-essential service, Pulse, to initiate attempts to request more user data, which is already the most aggressive move by a "big company" at present.

However, OpenAI's entry may not mean the end of the personal agent track for entrepreneurs. On the contrary, OpenAI's entry has opened up a market that was previously sealed by giants due to "caution" and has helped educate users.

Where are the opportunities for entrepreneurs?

Opportunities may lie in deep fulfillment in vertical fields.

Pulse's "non-essential" and "generalization" are precisely its weaknesses. An assistant that wants to do everything finds it hard to achieve extreme professionalism in every vertical field. This may be the breakthrough point for entrepreneurs.

In many specific life areas, such as fitness, diet, accounting, scheduling, note-taking, etc., one can enter a scenario where a user has clear needs and is most willing to "contract," to build a "fulfillment" capability that is ten times better than a general model. To become the "most knowledgeable and reliable" vertical agent still holds significance.

When you get the "trust-fulfillment flywheel" spinning at one point, you will have the most solid moat. I believe the future landscape may not be a centralized AI empire but a "trust federation" composed of countless trustworthy and professional vertical agents.

The emergence of Pulse can be understood as pointing a direction for the era of AI Agents. Opportunities still belong to those innovators who can deeply understand and solve the core problem of "trust-fulfillment" and ultimately create irreplaceable value for users.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。