This is about Instinct, a personal AI assistant that works via text, phone, or email like a human. Founder Noah Shinn believes all software will collapse into one simple interface, and Instinct is at the forefront—launched a year ago, zero marketing, 10% daily growth, and over $1 billion in annual transaction volume. He's bullish on travel and food delivery, arguing agents will reduce friction and boost transactions. Key holdings: Muse (Amazon's similar product but different approach), Uber Eats (agents will increase ordering frequency), and travel OTAs (50% of transactions from travel, hotels willing to pay 30% commission). Shinn rejects ads, earning via transaction fees.
Noah Shinn founded Instinct, a personal AI assistant that users can interact with via text, phone, or email, much like communicating with another person. Instinct has its own dedicated phone and computer, enabling it to perform nearly all online tasks on behalf of the user. Launched about a year ago
Noah Shinn is the founder of Instinct, a personal AI assistant with its own dedicated phone and computer that users can interact with via text, email, or phone calls, much like communicating with another person. The core theme of this issue is how personal agents are reshaping users' interaction with the digital world, and Instinct's unique path in trust-building, business model, and infrastructure. Noah Shinn believes that all software will ultimately collapse into a single simple interface, and Instinct stands at the forefront of this transformation — launched about a year ago, with zero marketing spend, growing approximately 10% daily, and already processing over $1 billion in annual transaction volume on the platform.
Noah Shinn argues that user trust in a personal agent needs to accumulate naturally over weeks, rather than being accelerated through forced authorization.
Shinn breaks down security concerns into two layers:
1. Storage security: This is a "solvable problem" that simply requires significant effort to ensure data isolation and lockdown.
2. New risk surface: For the first time, agents are empowered to autonomously use credit cards, access email, and calendars. Instinct has built a "firewall" system decoupled from the agent architecture — all content entering the agent is first intercepted/blocked/denied; every action the agent takes is paused, intercepted, and approved by an independent system before execution.
Noah Shinn explicitly rejects the advertising model, arguing that a personal agent should not act against user interests to influence their behavior.
Noah Shinn believes that personal agents will fundamentally reshape digital service industries, not through "disruption" but through "conversion" — reducing friction and increasing transaction volume.
Shinn advises traditional companies to respond to the agent era through "phased experimentation": Instead of full-scale openness, run A/B tests on 1% of users, observe changes in metrics like transaction volume and user satisfaction, then gradually expand.
Noah Shinn reveals that he spends about 40% of his time thinking about compute resources — the most critical bottleneck Instinct faces.
Noah Shinn believes the personal agent market is still in its early stages, and competition is not the current focus.
Noah Shinn believes that all software will ultimately collapse into a single simple interface, and Instinct is evolving in that direction.
| Position | Analyst View | Key Data |
|---|---|---|
| Muse (Amazon) | Neutral | "Good product, but different path"—Muse is a new app, Instinct focuses on simple access |
| Uber Eats / DoorDash | Bullish (agents will increase transaction volume) | Friction approaching zero will boost usage frequency |
| Travel Industry (OTA) | Bullish (agents will restructure) | 50% of transaction volume comes from travel; boutique hotels willing to pay 30% commission |
| Apple Pay / Amex | Comparable benchmarks | Business model reference: free for users, merchants pay fees |
| Shopify / Amazon / Apple | Commission rate benchmarks | 2.5-3% / ~10% / 30% |
1. Noah Shinn believes it takes about three weeks for users to trust an agent—40% of users share their credit card within three weeks, and once shared, retention reaches 80%. Trust is a "chicken-and-egg" problem: the more data an agent has, the more useful it becomes, but users should share at their own comfortable pace.
2. Shinn explicitly rejects the advertising model: "If you're not paying, you are the product. Instinct should not influence user behavior against their interests." He contrasts this with platforms like Google and TikTok, which offer free services in exchange for attention and then sell it to brands.
3. Shinn breaks down security issues into two layers: "storage security" and "new risk surfaces," the latter requiring an independent monitoring system decoupled from the agent architecture. Every action taken by the agent is paused, intercepted, and approved by an independent system before execution—"like a watchdog."
4. Shinn believes the agent era will "transform" rather than "disrupt" the digital services industry—reducing friction will increase transaction volume. He cites Uber Eats as an example: a user only needs to say "order what I usually eat," and the agent can complete the order, which will not reduce business but may actually increase it.
5. Shinn reveals that he spends about 40% of his time thinking about computing resource issues—the core contradiction is 10% daily user growth versus a 3–4 month lead time for computing resources. If calculated at a conservative growth rate, the user base could reach 100 million in 3–4 months.
6. Shinn puts forward the thesis that "all software will collapse into a single simple interface"—over the past 20 years, every time information needed to be displayed, a new app had to be built; in the future, this will disappear. Instinct currently has no standalone app; users interact through existing channels, and the long-term trend is "simpler."
7. Shinn believes agents will evolve from "task executors" to "goal pursuers"—users set high-level goals (e.g., "reach this pace within three months"), and the agent autonomously plans and executes. Some users already run their entire small business backend on Instinct.
8. Shinn describes Instinct's invite-only growth as "the strongest word-of-mouth growth case I have ever seen"—zero marketing spend, 10% daily growth, with users even selling invites for $300 on eBay. Each user has only five invite slots, but about 10% of users choose to use one each day.