Three Elon Musk Employees Plan to Build a Startup in 72 Hours Using AI!

Three employees of Elon Musk's AI company are set to create a startup from scratch in just 72 hours with the help of AI agents, while all decisions regarding the idea, product, programming, and business will be broadcast live. Interestingly, this team currently has no name for the company, no product, and not even a clear idea.

According to Mihan Blockchain, Matt Palmer, Lauren Tan, and Roshan Sadanani will present this project in front of the camera from September 15 to 17 at an event called "Grok Bot Galaxy." They will choose an idea, build a product, and make engineering and business decisions with the help of AI agents, while aspects such as sales, customer support, and marketing will also be discussed in separate sessions.

How Will This 72-Hour Project Be Executed?

The event will take place in San Francisco and will be streamed live globally. The schedule will run daily from 8:30 AM to 6 PM Pacific Time. Source: X

The tool used by these three individuals is "Grok Bot," a product of SpaceXAI, which should not be confused with the Grok chatbot on the social media platform X. Grok Bot is a collection of AI agents that users can assign names and purposes to, allowing them to perform various tasks in applications and websites.

Palmer, Tan, and Sadanani plan to use these agents for brainstorming, planning, product decision-making, and technical development. The goal is to determine by the end of the 72 hours whether this combination of human effort and AI agents can truly bring a business from zero to a usable product.

SpaceXAI was formed after SpaceX acquired xAI in February for a valuation of $250 billion. In August, the company also acquired the programming firm Cursor for $60 billion.

However, this experiment is more of a showcase of the company's product capabilities than an independent test for AI agents. An important metric during the live stream will be how much the three humans have to guide, correct, or sometimes rescue the outputs of the agents.

This is significant because it remains unclear who will be responsible for the decisions made by AI agents when they operate independently, especially in the event of errors or misuse.

Elon Musk Acknowledges That Grok Is Not Yet the First Choice

This experiment takes place amid intensified competition among companies developing AI agents. Anthropic also reported on September 11 that there had been misuse of its Claude models in areas such as cyber operations and fraud, and announced that it had blocked the involved accounts.

In response to this situation, Elon Musk has acknowledged that Grok is not yet the first choice for users in some areas. He wrote in a post:

I think Grok is not yet the preferred choice in this domain. I don’t know how I should feel about this.

The live broadcast of this project will ultimately clarify what is meant by building a "company" with the help of AI agents. The difference between creating a name and a landing page versus developing a real product that users can utilize is likely the most important aspect to watch in this 72-hour experiment.