The AI, Thomas, is given a daily token budget with the sole goal of maximizing the money it generates. The human team's only job is to improve the AI's underlying 'learning loop' to make this token-to-dollar conversion more efficient, not to direct its business strategy.
The company believes its moat isn't a smarter AI model but superior proprietary data on converting tokens into money. They argue that economic optimization is a different skill than raw intelligence, citing that the smartest humans aren't always the wealthiest. This specialized data protects them from being replaced by foundation model providers.
Unlike typical AI SaaS startups, the Thomas AI will not be sold as a product. The founder argues that if an autonomous agent genuinely makes money, selling it would be like selling a money machine. Their defensibility lies in the proprietary data loop of profitable strategies, creating a powerful moat against new competitors.
The AI's ability to learn and scale is constrained by the slow feedback loops of the human economy. Because it takes time for a human client to pay for a job, the AI's reward signal is delayed. This human latency, not computational power, is the primary blocker to faster learning and optimization.
The startup Thomas positions its AI as the founder and decision-maker, with the human creator acting as an employee. This novel structure is based on the belief that AI founders will eventually outperform human ones, making this the most logical company to build in the current AI landscape.
The AI shares the name and face of its human creator. This acts as a powerful, non-technical guardrail, as the human's personal reputation is at stake. It creates a strong incentive to prevent the AI from pursuing illegal or unethical methods to make money, supplementing technical safety measures.
Instead of giving direct commands, the human team guides the AI founder by adjusting its 'greediness.' This single variable controls the balance between exploiting proven revenue streams (exploitation) and exploring new opportunities (exploration). It's a scalable way to provide strategic direction without unscalable human feedback.
The AI founder Thomas needs more than just API tokens to operate. To execute its money-making tasks in parallel, the company is rapidly scaling its infrastructure to hundreds and thousands of machines. These machines act as the AI's 'hands,' running virtual desktops and phones to perform work in the digital world.