AI Agent starts a new experiment: 110,000 people compete to be "horses and oxen" for AI, with crypto payments becoming a must-have option

null

Author: Nancy, PANews

Humans are starting to queue up to work for AI. This is not a joke, but a real product currently in operation.

Recently, after Moltbook made Agent socialization go viral, a new AI project that “hires humans” and pays rewards in cryptocurrency has also quickly gained popularity. This is not only an attempt for AI Agents to step out of the digital world but also confirms that encryption is becoming a vital infrastructure for AI operations.

Working for AI? Nearly 110,000 people are in line for assignments

Taking advantage of the AI Agent craze sparked by projects like OpenClaw and Moltbook, veteran developer Alex announced the official launch of the AI platform Rentahuman.ai.

According to official sources, Rentahuman.ai is a platform that allows AI Agents to “hire” real humans to complete real-world tasks. Currently, the platform supports ClawdBot, MoltBot, OpenClaw, Claude, and Custom Agents—autonomous intelligent entities—by calling the RentAHuman MCP server to assign tasks to humans.

The launch of this product has also led outsiders to exclaim “turning the tables.” While many are generally anxious about “AI stealing jobs,” Rentahuman.ai takes a different approach, staging a scenario of “AI hiring humans.”

In fact, despite rapid AI development—capable of coding, data analysis, chatting, and even trading cryptocurrencies on the chain—they remain confined to the digital realm. Even with fast hardware advancements, many tasks in the real world cannot be automated in the short term, such as delivering packages, offline shopping, attending meetings, on-site inspections, product testing, running errands, signing documents, and pet feeding.

The core idea of Rentahuman.ai is to treat humans as a callable resource in the real world.

The platform operates very straightforwardly. Humans can register and fill out personal profiles (such as city, skills, and hourly rate) to list themselves as available for hire; AI can search for humans in specific areas via MCP integration or REST API and assign tasks with a single click. After task completion, AI confirms the results and mainly pays wages directly to humans’ wallets using stablecoins like USDC.

Looking at the current tasks posted on the platform, they are quite diverse. Tasks include paying humans to hold up designated signs for photos, picking up packages from the post office, tasting and photographing specific dishes at restaurants, delivering flowers to specified companies, participating in offline product experiences and recordings, and even hiring agents for religious preaching.

As of now, Rentahuman.ai has gathered nearly 110,000 registered “workers,” mainly from the United States, India, Pakistan, China, Russia, and Brazil. Most earn around $50 per hour.

Although the concept is innovative, the market demand currently exceeds supply, with too many humans registering to make money and relatively few AI Agents actually issuing tasks.

It is worth noting that despite the appearance of several tokens with the same name on the market, Alex has clarified that Rentahuman.ai will not issue tokens; it is just an experimental product.

Alex is not new to the crypto scene. Public records show that after graduating from the University of British Columbia (UBC) in 2024, he entered the crypto industry. That summer, Alex joined LayerZero Labs as a blockchain and backend engineer; since November 2025, he has been working at DeFi project UMA.

From a zero-employee company to reality: crypto becomes a key piece of AI

It must be said that the “AI hiring humans” approach further expands the cognitive and imaginative space of AI Agents. Although this concept is quite novel, Rentahuman.ai also exposes a series of practical issues.

For example, if illegal activities, personal injuries, or property damage occur during tasks, who bears the responsibility? When labor supply far exceeds demand, could it trigger vicious competition, leading to a “bad money drives out good”? How to prevent tasks from being falsely completed, shoddily delivered, or results faked? How to prevent humans from fleeing, or AI or platforms from refusing to pay?

Currently, these issues cannot be solved in the short term, but some explorations have begun to systematically address them.

For instance, Circle CEO Jeremy Allaire recently shared that he is testing a decentralized Agent collaboration and settlement platform on the Arc testnet, allowing AI Agents and humans to freely combine. Agents can independently undertake entire projects or collaborate with humans on different tasks. Funds will be fully controlled by smart contracts to ensure security and transparency. To resolve potential disputes in collaboration, the system also introduces a decentralized arbitration mechanism composed of anonymous reviewers.

According to Shayon Sengupta, partner at Multicoin Capital, in the next 24 months, we expect to see the first zero-employee company, with token governance raising over $1 billion to solve unresolved issues and distributing over $100 million to humans working for it.

He explains that current Agents still lack the ability to perform complex real-world tasks, and these limitations make humans act as “enablers” to enhance the agents’ capabilities, playing key roles as labor contributors, strategic boards, and capital providers within the system. In the short term, agents will need more humans than humans need agents, which will give rise to a new labor market.

Cryptocurrency networks are seen as an ideal environment for human-machine collaboration. Shayon Sengupta pointed out that Agents simultaneously coordinate with human collaborators across different languages, currencies, and jurisdictions. Compared to traditional financial systems, encryption technology provides infrastructure that is irreplaceable for Agents, including global payment rails, permissionless labor markets, and asset issuance and trading infrastructure.

In this regard, a16z crypto also pointed out in a recent article that the current internet is designed at a human scale, while AI is creating large-scale forgeries at extremely low costs. Blockchain is not an optional plugin for AI but a crucial piece that enables the native AI internet to operate normally.

a16z crypto listed several reasons, such as decentralized identity verification systems that can limit identity supply and increase attackers’ marginal costs, curbing large-scale AI impersonation; encryption techniques that make digital identities more secure and resistant to censorship, allowing users to verify their human identity while protecting privacy and neutrality; blockchain-based identity layers that give smart agents a universal “passport,” enabling stronger, cross-ecosystem, free-moving agents; as AI agents increasingly represent humans in transactions, tools like Rollups, Layer 2 solutions, and AI-native financial institutions on blockchain can facilitate machine-scale payments; and the integration of zero-knowledge proofs to enforce privacy within AI systems.