Meet Noah Shinn: The 23-year-old college dropout whose viral AI assistant is emerging as a new challenger to Mark Zuckerberg’s Meta

Noah Shinn is 23, dropped out of college, and has turned a private AI assistant into one of the hottest new companies in the field. Instinct, the startup he founded last year, raised $250 million at a $2.5 billion valuation in August, then added another $1 billion in a Series C this week, pushing the company to $10 billion. That is a fast climb by any standard. It also puts a sharp price tag on the current obsession with AI agents.

Instinct is still only open to invited users, yet Shinn said on the Invest Like the Best podcast that it is approaching $1 billion in annual transaction volume. He also said the company has spent nothing on marketing and has grown mostly through word of mouth. The demand has gotten odd enough that existing users are selling invitations on eBay. The assistant itself is free, and it can talk to users by message or phone call.

The product first launched in private beta in February, according to The Wall Street Journal, and last month it moved to an invite-only beta with five invites per user. Shinn has said people are using it for things like planning weddings and canceling subscriptions. Travel appears to be a major use case: he said that category makes up half of the yearly transaction volume flowing through the platform. His pitch is that the assistant learns preferences, stores them in memory, and uses that memory to make future bookings easier.

Meta has noticed. Alexandr Wang, Meta’s chief AI officer, has spent the past week mocking Instinct on X after Meta announced its own personal AI agent, Muse. Wang compared the two in jokes about MMA, speed, and even the spelling of Instinct’s name. Shinn said he spends very little time worrying about competitors, but the posts may have helped anyway; they drew hundreds of thousands of views and more than 1,000 likes.

Shinn’s path to this moment started well before Instinct. He grew up in the San Francisco Bay Area, attended Northeastern University, and left in 2023. As a student, he worked on type prediction, code generation, molecular dynamics, and aerospace projects, then coauthored the paper on Reflexion, which used memory and feedback to help language agents improve. That work was cited hundreds of times and presented at NeurIPS in 2023. He later joined Sierra as one of its first employees before striking out on his own.