The 'Roller Effect' of AI and the 'Workshopization' of Software as a Prelude to the Internet of Intelligent Agents

Author: Meng Yan's Blockchain Thoughts

Some interesting things have happened in the past month or so.

On August 11, SpaceX released the AI agent application Grok Bot. This is an "AI colleague" that works around the clock on its own cloud computer.

On September 8, Meta launched the personal agent application Muse. It can open browsers, fill out forms, and negotiate on behalf of users. Within ten days of its launch, it topped the US App Store charts and triggered a sell-off of traditional online intermediary platform companies, such as Expedia, Airbnb, and Booking, which saw their stocks plummet.

On September 22, Anthropic, the company behind Claude, released a new model, Opus 5.5. Not long ago, OpenAI, the creator of ChatGPT, launched two new versions of GPT-6, both priced at half of their predecessors.

In mid-September, one of the world's largest SaaS software companies, Salesforce, held its annual conference. Patrick Stokes, the president responsible for application business, stated that AI would dismantle software interfaces and then replace them.

These events come from different companies and industries, concentrated within two months. They may seem like separate blows, but underneath is the same story: AI agents capable of handling tasks for people are beginning to interact with software and the internet. The shape of the new generation of the internet is emerging.

I. The Roller Effect

A head of an AI incubator told me that the biggest awakening for AI entrepreneurs over the past year is realizing that AI is fundamentally different from the internet and blockchain; it is an arena of extreme centralization and rapid centralization, where the lifespan of startups is very short. They must seek speed and sales, finding buyers without hesitation before being crushed by giants, cashing out and exiting. As for dreams of growing their own businesses, they shouldn't even think about it.

This is the mindset of the small grass in front of the roller.

The "Roller Effect" is a source of anxiety for many AI entrepreneurs. With each advancement in cutting-edge models, a batch of startups is unknowingly crushed. Publicly traded companies can at least mourn their stock price plummets, but the instant zeroing out of hundreds or thousands of small teams goes unnoticed.

First, let's look at the money. According to US venture capital data firm PitchBook, global AI venture capital reached a record in the first half of 2026, with over half flowing to OpenAI and Anthropic. This is like a banquet with thousands of tables, where half the dishes are served at the main table, which only has two people sitting at it.

Next, let's look at specific companies. Google has an AI note-taking product called NotebookLM, where users can input documents and web pages, and it transforms the content into podcast-style audio explanations, set to become popular online in 2024. Its head, Raiza Martin, left Google with two colleagues to create an AI podcast application called Huxe, which generates daily audio briefings based on users' emails and calendars, with funding from Google's chief scientist Jeff Dean.

On May 21 of this year, Spotify, the world's largest music streaming platform, made an update that included similar features. On May 22, Huxe announced its shutdown.

Publicly traded companies are no different. In February, Anthropic released a set of plugins for Claude in industries such as law and finance. Thomson Reuters, the parent company of Reuters, relies on selling legal, tax, and accounting information, and its stock price dropped by over 15% that day.

After Claude released the Opus 5.5 model, a large number of users posted beautifully crafted videos on Twitter made with simple prompts, celebrating, while teams that had spent countless sleepless nights in the video AI production field could only face their screens and cry silently.

The world is indifferent, treating all things as mere fodder. Giants like OpenAI and Anthropic do not intend to go against you; they have no hostility towards you and no intention to compete with you, and they may not even be aware of your existence. They simply move forward with the trend, and you vanish into thin air.

What did that cliché say? Eliminating you has nothing to do with you.

II. The Workshopization of the Software Economy

In February of this year, Claude Opus 4.6 was released. With the support of this model, Claude Code suddenly became so powerful that even those who cannot write code felt that their technical dreams could be salvaged.

More and more people are getting into software development. There is an AI platform called Lovable that allows non-programmers to create websites and applications through chat. It claims that the platform adds one million new projects each week, with users primarily being programming novices who can't write a single line of code.

On the other hand, selling software has become challenging. The main index fund tracking US software stocks fell by over 24% in the first quarter of this year, marking the worst quarter since the 2008 financial crisis. Stock prices are based on expectations, and while high interest rates are a factor, it does not mean demand has shrunk. In September, the internet intermediaries took the hit, and the market is repricing software sold to consumers.

I call this the "workshopization" of the software economy. The industrial revolution moved spinning and weaving from homes to factories; AI is moving software writing from factories back to homes. Every individual and small company can set up their own software workshop.

In the past, opening a small store required a complete inventory system, which meant negotiating with outsourcing companies for requirements, signing contracts, and waiting for construction periods. Now, you can just say a few words in front of a computer, and if you're quick, you can have a prototype ready in an afternoon. Internal tools like customer management and personnel scheduling are the same.

Can the next big company grow out of these workshops? It's difficult. The problem with workshops is isolation, which becomes more apparent as the number of workshops increases. In a small country with few people, the sounds of chickens and dogs can be heard, but the people do not interact until they die. Laozi regarded this as an ideal, but in software, it becomes a dilemma: I can't use your product, and you can't touch my data; the things made in workshops can only be used by themselves.

There have never been so many people making software, and there has never been such difficulty in selling software.

III. The Internet of Intelligent Agents

On the AI side, there is extreme centralization, while on the software side, it is moving towards decentralization. It is a tale of two extremes, but in fact, they are two sides of the same coin. It is precisely because your AI capabilities are continuously strengthening that the space for software is being squeezed.

This is just the beginning. People are still using AI to develop software and then using that software. As we move forward a few more steps, software may no longer be needed; users will interact directly with AI agents to solve all problems.

Muse is currently the clearest prototype.

Meta allocates a virtual computer to each user in the cloud, and Muse works on this computer. When users close the app, it continues to work, notifying users only when it has completed tasks or requires approval; if there are interfaces available, it directly calls them through connectors; if there are no interfaces, it opens a browser and operates page by page like a human.

Filling out forms, comparing prices, negotiating—tasks that used to require user involvement are now handed over to it.

It still doesn't do well. The US payment media PYMNTS asked it to accomplish three tasks: restock toilet paper on Amazon, order a Domino's pizza, and reserve a table at a restaurant. It failed to complete any of them, with the evaluator stating that it added an extra layer of management to tasks that a person could finish in 30 seconds.

But what we need to look at is the direction. How is this different from using apps in the past? In the past, to book a flight, you had to open several travel websites, compare prices one by one, and fill in names and ID numbers in various fields. In the future, you can just tell Muse, "I want to go to Tokyo next month, find the cheapest direct flight, and handle the rest."

For the first time, users will only speak to one agent, which will interact with the entire internet on their behalf. Apps and websites will retreat to the background.

This is not just a change in how people interact with the internet; it is another upgrade of the internet itself.

Looking at the history of the internet: Web 1.0 connected documents; it was the internet of documents, where people read. Web 2.0 connected applications and services; it was the internet of applications, where people used. What does Web 3.0 connect?

The term Web 3.0 has had several interpretations over the past twenty years. Tim Berners-Lee, the inventor of the World Wide Web, systematically described the "Semantic Web" in 2001, aiming to label web pages with tags that machines can understand. The blockchain community proposed the "Value Internet," aiming to allow money and assets to flow online like information. Both descriptions highlight characteristics but do not specify the form.

Now it is clear that the form of Web 3.0 is the internet of intelligent agents. Agents read, use, and negotiate on behalf of people, with one agent interacting with hundreds of apps and thousands of APIs. Human experience is no longer important; what matters is whether the agent finds it good, and if it is good for the agent, it is good for me. For the past twenty years, product managers and designers have focused on human experience in product development, competing on interfaces and operations. In the future, products will need to be designed for agents, competing on clear interfaces, clean data, and readable terms.

Salesforce launched a toolkit without interfaces, opening the entire platform to agents through MCP and APIs, and directly integrating its own functions into Claude. A company that sells software through interfaces has dismantled its own interface.

The CEO of online travel platform Expedia summarized the company's new strategy as needing to "appear wherever agents are." In the past, travel platforms needed to attract people to their websites; now they need to go to the agents.

Who is this bad news for? Wall Street is repricing "consumer inertia." Many businesses rely on customers being too lazy to compare prices, switch apps, or negotiate over the phone. Agents do not mind the hassle; they will handle all these tasks. Therefore, the market is selling off such businesses.

Even Meta itself needs to transform. This company started by selling user attention, with revenue primarily coming from advertising. Agents do not view ads or scroll through information feeds; Mark Zuckerberg's revenue source for Muse is a small fee from transactions.

Thirty years ago, websites began optimizing for search engines, which later became a business. This time, the object of adaptation has shifted from a sorting machine to an agent that makes decisions on behalf of its owner. Products that agents cannot understand are equivalent to non-existence; products that agents can understand and use smoothly will be able to secure business. The capabilities produced in workshops are the same; if agents cannot understand them, they cannot be sold.

US tech analyst Ben Thompson stated that agents will become the "ultimate gatekeepers." Whoever controls the agents controls user demand.

Another inevitable development is blockchain payments and token economies. Agents exchange value among themselves, with small amounts and frequent transactions, often with unfamiliar counterparts. Whether to pay and how much is decided by programs on the spot, with no human oversight. Credit cards charge 2.9% plus 30 cents per transaction, making a 1-cent payment incur 30 times the fee; card networks cannot handle such transactions.

Visa's own research also acknowledges that cards cannot handle this segment, assigning micro-payments between machines to stablecoins. Stablecoins are digital currencies typically pegged to the US dollar on a one-to-one basis, circulating on the blockchain, with a total supply exceeding $300 billion.

My judgment is that blockchain-based digital payments and token economies, which involve using on-chain credentials for pricing, settlement, and profit distribution, are essential elements of the internet of intelligent agents.

I remember a meeting at the Digital Asset Research Institute in 2018, where Professor Zhu Jiaming said that in the long run, blockchain is not for people but for AI. At the time, I felt this was correct but couldn't understand how AI would use it. Now it is clear.

IV. AI Agent-Friendly Jobs

The impact of AI on jobs has always been a hot topic. What kind of people can survive and develop sustainably in the AI era? Or to put it more bluntly, what jobs can escape being crushed by the roller?

They must be jobs that collaborate well with AI agents.

Top AI scientists in large model companies certainly fit this description, but such individuals are extremely rare and have a high barrier to entry, making them inaccessible to most people.

Recently, the trendy role of Frontend Deployment Engineer (FDE) seems to have quickly fizzled out. People soon realized that this was merely a fashionable term for on-site outsourcing, and the key issue was that it often turned out to be a one-time job. Various imagined AI enterprise solutions rushed in, only to find that the data quality was too poor, and the deployed enterprise AI was as clueless as a village fool, forcing them to start reworking from the bottom data. But isn't this the job of data engineers? As a result, FDE became Fooled Data Engineer, a data engineer fooled by data. Who wants to do this lousy job?

Currently, the hot roles are AI engineers who develop products based on AI models and AI-Assistant Developers who use AI to assist in traditional software development. However, if the judgments that AI agents are squeezing software and the internet towards intelligence are correct, then these two job roles will also transform. In the past, products were made for people; in the future, more and more will be made for agents. In the past, optimization was for search engines; in the future, it will be for agents.

Another more common role is the management of organizations of AI agents.

Once agents reach a certain level of capability, there will be no need to develop software; you just need to manage them well, make demands of them, and make judgments and decisions at critical junctures to accomplish the vast majority of tasks. But that is still far from enough; in the future, you will need to design and organize dozens, hundreds, or even thousands of agents, turning them into an efficient and powerful legion, designing effective processes, ensuring safety, controlling expenses and token budgets, effectively evaluating performance, and continuously improving, then competing and battling against equally scaled agent legions of your competitors.

For example, a company's procurement agent needs to negotiate prices with sales agents from a thousand suppliers, and the opposing agents are equally intelligent. The outcome depends on how this side divides labor, authorizes, and holds accountable. Management must become engineering, with permissions, boundaries, incentives, and audits all needing to be trainable and measurable. This is a new management science and a new system engineering.

Conway's Law states that the organization that designs a system will ultimately produce a system that replicates its own communication structure. In the agent era, the members of the organization are agents, and the system and organization become one. Whatever the organization looks like, the agent system will look like; designing the agent system is designing the organization.

This is not just someone else's fantasy. Microsoft stated in its annual workplace trends report that everyone will become the boss of agents; NVIDIA CEO Jensen Huang said that IT departments will become the HR departments for AI agents.

Therefore, in the internet of intelligent agents, the most valuable jobs will be AI engineers who can create products and services for agents, and commanders who can lead a thousand agents to defeat another thousand agents.