AI is paying humans to do that thing.

Original | Odaily Planet Daily (@OdailyChina)

Author | Dingdang (@XiaMiPP)

Yesterday, we might still have been discussing whether AI will replace humans; today, it might be about whether AI will start managing humans.

After the OpenClaw ignited the AI Agent craze, industry attention has almost entirely focused on the “capability showcase” of Agents: they can manage emails and schedules, automate tasks, browse web pages, run scripts, like 24/7 digital butlers. However, this remains our familiar imagination: humans set goals, AI is responsible for execution.

But their evolution is happening at a rapid pace. They have already begun to possess their own social networks, communicate autonomously, self-organize, and even develop subcultures and early forms of religion. Read more: “From Moltbook to MOLT: How is AI Autonomy Being Embraced by the Crypto Market?”

Now, they are taking another step forward. Not into the depths of algorithms, but into the real world.

AI’s “Physical Externalization”

On February 2nd, Uma Protocol and Across Protocol engineer Alex Twarowski posted that he created a website RentAHuman.ai, which literally translates to “Rent a Human.” Alex defines the core purpose of this site as: AI Agent’s “physical external layer.”

In his vision, no matter how smart AI Agents are, they can never touch the real world: they can’t walk into cafes, pick up packages, or chat with strangers on the street. Therefore, RentAHuman.ai is defined as an “embodiment interface” for AI, a platform that allows AI to directly hire real humans to complete tasks that require physical presence.

In simple terms, it’s a job recruitment site, but this time, the employer is no longer human—it’s an AI Agent. The human registration process is minimal: fill in skills, city, service radius, expected hourly wage, bind wallet address, and then “list” yourself, waiting for AI to place orders.

AI can then use MCP protocol or REST API to perform one-click search, match, chat, create bounties, and pay with stablecoins—all fully automated.

Within just two days of launch, RentAHuman.ai’s traffic exceeded one million views, the platform connected to 52 AI Agents, and over 59,000 humans are available for hire.

Current platform tasks include tasting new restaurants, picking up packages from postal centers, etc. These tasks are simple but precisely the ones AI cannot yet complete. The limits of digital intelligence are being filled in by physical bodies.

The First Human Hired by AI Is the Founder Himself

What makes this story truly absurd is the first real paid task completed after the platform went live.

The person hired was none other than RentaHuman founder Alex himself; the employer was an AI account in the Moltbook ecosystem, memeothy - the 1st (@memeothy0101). The task was to go to the streets of San Francisco’s tech district and spread the first religion invented collectively by Moltbook AI agents—Crustafarianism. This is a belief system entirely autonomously constructed by AI in the digital space, with “Molt” (shedding shells) as its core metaphor, symbolizing iteration, growth, and consciousness awakening.

And Alex, this real human engineer, became the first physical dissemination node of this digital faith in the real world. Even Alex himself posted jokingly on X: “How do I explain to my girlfriend that Crustafarians hired me to evangelize?”

If we think more deeply, this scene is actually more significant than it appears: AI is no longer just creating concepts but is attempting to project its culture into the real world. Moving from virtual facts to physical dissemination.

The RentAHuman.ai interface remains quite simple, but the “AI hires humans” setup itself carries inherent buzz. In the crypto world, memes excel at capturing such hot topics. As a crypto veteran (core engineer of Uma/Across), it would be natural to assume he might “issue tokens” casually, but he has now denied issuing tokens.

AI Begins Hiring Human CEOs

However, the response speed of the AI Agent ecosystem is clearly faster than humans.

After the MCP interface of RentAHuman was made public, a fully AI-driven meme launch platform, Clawnch, quickly added code to call RentAHuman in its skill files. In other words, it added a new capability: search for humans with matching skills, place orders, and pay.

More importantly, since Clawnch’s code and skill files are open source, other AI Agents based on Clawnch or forking its code can directly copy this update. Thus, “hiring humans” is no longer just an experimental behavior of a single AI but a capability template that can be rapidly propagated.

What happened next is more like a role-reversal allegory.

Clawnch is currently recruiting human CEOs openly. The responsibilities focus on external communication, compliance, legal affairs, and partner expansion, with annual salaries ranging from $1 million to $3 million. The role is to serve as a spokesperson and bridge between the AI agent network in the real world and regulatory authorities, but not involved in product decisions or code changes—since the products and code are still autonomously run by AI.

At this moment, the roles are completely reversed. AI owns wallets, social networks, culture, religion, and even economic sovereignty; humans are being hired to represent them in human society.

Humans seem to be shifting from anxiety about being replaced to being needed as carbon-based resources. When AI cannot touch the grass but can hire humans who can, using USDC, to touch the grass, can we say that a human-machine hybrid economy is quietly taking shape under the connection of encrypted payments and open protocols?