Robinhood CEO Says ‘Everyone’ Should Use This Method to Work While They Sleep

Forget doing everything yourself. Robinhood CEO Vlad Tenev envisions teams of AI agents that keep building and testing through the night while their human managers sleep. 

On the latest episode of the Sources podcast, Tenev said that his “North Star” is for everyone to have “a team of agents” to be productive while they rest. 

“So we should be a 24/7 company,” he said. “Each person should be a 24/7 agent manager — your job becomes managing your agents.”

AI agents are software systems designed to work toward a goal without needing human involvement for every step. They can break down an assignment into smaller tasks and figure out the best way to tackle them.

They also use tools like web browsers to fulfill work tasks. Some can also spot mistakes and assess their progress. 

Workers use agents for tasks like gathering information from multiple sources and turning it into a research report. They can also use them to book travel, build apps and websites, manage emails and analyze stocks, in addition to other jobs.

Tenev said he envisions a world where AI agents generate “a large portion” of the trading volume for individual investors. Each person ensures that their AI agents have the context to carry out their tasks. 

Tenev said AI agents have advantages over human traders: They can work around the clock, buy and sell across different asset classes simultaneously and navigate Robinhood’s trading options with ease. They “don’t have a lot of the limitations that we have as human traders,” he said. 

Why AI agents need to test their own work

Tenev’s vision extends beyond investing. On the podcast, he said using AI to write code had turned him into an engineer again, and he expects the technology to reshape how companies build products. 

However, keeping AI agents productive overnight requires more than handing them an assignment. They need enough information to understand the job and a way to check their results. 

“They have to be able to test their work,” Tenev said. Without that capability, he explained, an agent eventually stops and hands the task back to a person.

That means companies must rethink the systems behind software development, including how they test code and release updates. Tenev envisions a continuous cycle in which agents build, evaluate and refine their work, with less need for people to intervene at each step.

“You’re going to see an end-to-end automation of product development and infrastructure development,” he said.

Tenev said users can easily automate research with AI. Researchers already have structured experiments and automated evaluations for testing models. 

“People thought that would be the hardest and the last to go because it’s just complex work,” Tenev said.