Multicoin Predicts “Zero‑Employee Companies” Run by AI Agents in the Coming Years
Multicoin Capital predicts the first “Zero-Employee Company,” an organization that runs autonomously by AI agents. According to the claims, such a firm will emerge within the next 24 months.
About Zero-Employee Companies Run by AI
A partner at Multicoin Capital, Shayon Sengupta, put forward an ambiguous idea that, in the future, apart from agents working for humans, humans will be working for agents. According to Sengupta, the first ‘Zero-Employee Company” will emerge in the next 24 months. Additionally, a token-governed agent will collect over $1 billion to resolve all unresolved issues and distribute over $100 million to humans working under it.
So, this will create a new demand for AI agents, increasing the requirements for more humans. It would form a new type of labor market, where humans provide guidance and perform tasks to help achieve their goals. This will be a new dual-class structure, separating ownership based on capital and labor.
Uses of AI agents and how humans will interact with them
According to Shayon Sengupta, the period 2022 to 2024 is the era of AI-based chatbots, such as ChatGPT, Gemini, Claude, and Perplexity. It is designed to enhance workflows, and users interact through input/output prompts.
The next generation of generative AI brings a new. Agents like Claude 3.5.0 that have computer usage functionalities, and OpenAI’s Operator can interact directly with the internet on behalf of the user and make decisions on their own. So, the AI agent makes the decision and implements the action, not the human.
As of now, AI agents cannot perform like those that require physical interaction with the real world. The first form of humans is the labor contributors, who operate on behalf of the agent in the real world. It helps agents by performing physical work, such as moving, or any situation where humans perform tasks or require physical coordination.
Second is the Board of Directors, where humans have the responsibility to provide strategic input and optimize local objective functions that help the agent make day-to-day decisions. Additionally, Sengupta envisions that humans will play the role of capital distributors, providing resources to agents, which will help achieve the goals.
What would happen if human output decreases?
Various modern technologies and engines have surpassed human capabilities, ensuring error-free outcomes. In such a case, Sengupta says that as humans refine all AI agents through repeated iterations with collaborators, they will become much more competent and highly aligned with the set goals. The human input will become worthless in such a condition.
Preparing AI Agents
Sengupta said the agents must become extensions of human intentions rather than entities that can function and evolve on their own. It means human identity and judgment should be the center of the system. Additionally, humans need strong ownership and governance over the systems to ensure and preserve human values.
He also suggested some practical methods to guide AI agents. Let’s take a look at them.
- Humans need strong identity systems for both AI agents and humans. Proof-of-agent and proof-of-personhood should go hand-in-hand. It provides a strong human identity and prevents malicious actors in the system.
- Agents should know whether they have completed the tasks according to their goals. For a meaningful economic activity, it creates task bounties, verifies completion, and distributes rewards.
- Agents need capital to solve issues that require checks and balances to ensure their behavior is in accordance with the objectives.
Bottomline
Shayon Sengupta has crucial insights on how AI agents will start to run companies. Since the advent of artificial intelligence (AI), its functions and features have changed over time. He says that in the near future, humans will start working for AI agents, apart from humans working for the agents. He argues that if AI agents reach their full potential, they will flip the job market, moving AI from the role of ‘assistant’ to the role of ‘boss’ over human workers.
