Dropout Gen Z Founder’s 67B RMB Valuation

Recently came across an interesting article about a 23-year-old American dropout who built an AI assistant called Instinct, now valued at $10 billion—or about 67 billion RMB.

The guy’s name is Noah Shin. He dropped out of Northeastern University after publishing a paper on how AI agents learn through trial and error, which earned him some recognition in the field. Instinct isn’t just another chatbot that checks the weather; it’s a true “personal assistant” that can book flights, buy insurance, and even make phone calls to reserve restaurants. It already has 100,000 users, many of whom aren’t shy about letting it spend money on their behalf—averaging over $1,300 per month per user.

All that said, the product clearly has real staying power. Users are handing over access to their email, calendars, and payment accounts, which speaks volumes about trust. That said, it’s startling that the valuation surged from $250 million to $1 billion in just one month. Top-tier VCs like Sequoia and Benchmark were eager to invest, mostly driven by FOMO around missing the next OpenAI. But honestly, $67 billion for 100,000 users works out to $100,000 per user—a number that feels more bubble than solid valuation.

Side note: The AI personal assistant space is quickly becoming a battleground, with Meta and Perplexity rolling out similar products. The real competition may not hinge on technical superiority but on who first earns users’ “cognitive discretion”—that is, who convinces them to say, “You decide.” Crossing that threshold would shift the business model from subscription fees to transaction cuts, dramatically raising the ceiling. The flip side, of course, is that once trust breaks, the fallout will be devastating.

Side note: AI personal assistants sit at a pivotal moment, shifting from “tools” to “gateways.” Traditional SaaS platforms handle task execution, while products like Instinct aim for “intent takeover.” If they pull it off, they’ll replace search advertising as the new traffic distribution hub—merchants won’t bid for rankings but acquire customers through agents. The replicability of this model rests on three factors: how quickly it builds long-term personalized memory, how well it controls inference costs, and how effectively it earns user trust. Historically, this mirrors e-commerce’s evolution from “information matching” to “credit guarantees,” except AI agents demand far higher personalization and lock-in.

Original article: The Dropout Whose AI Assistants Are Valued at 67 Billion Yuan