Instinct， Operated by Spear Street Technology, founded by Noah Shinn, currently around 23 years old, had a conversation with Patrick O'Shaughnessy on the podcast Invest Like the Best on September 28th. Just a few days ago, he completed a $1 billion financing, and the valuation has increased from $1 billion a month ago to $10 billion. In this interview, we can see how he started doing Instinct, with a lengthy breakdown of how he developed products, operations, growth, business models, and competition. It is a good material for first-hand learning. You can take a good look on the way during the holiday. The main content inside is: growth Zero marketing, reputation growth. He said that the user base has been growing by about 10% -11% per day recently. The annualized trading volume (GMV) of the platform has exceeded approximately $1 billion, of which about half comes from travel. Product aspect In the early stages, we should pursue "comprehensibility" instead of piling up abilities: users need to be able to anticipate what it will do and whether the results will be natural. He believes this is one of the reasons why the reputation exceeded expectations. Social intelligence, compared to previous assistants, some reminders do not disturb users very frequently, which is because their agents have some social intelligence technology solutions. Booking is very important, they are developing a new booking system and addressing the new features of bidirectional demand. We will accurately match the needs of both parties. For example, if the user says, 'This is very important, it's my spouse's 30th birthday.' The restaurant says, 'Priority should be given to birthdays, especially those around the 10th birthday or important anniversaries.' Describing the requirements and explaining why important logistical burdens are no longer needed, both parties can match accurately. Users are more likely to book seats at important moments, and restaurants also attract a better customer base. The agent will record user preferences and only tell them once, and will remember your preferences next time. On Trust After about three weeks of use, about 40% of users will hand over their personal credit cards to it; About 80% of users who have shared at least one sensitive information are retained. Authorization should be gradually released according to the user's pace and can be revoked at any time. Business Model Tend to make users free and take rates from merchants, rather than advertising. The reason is that assistants who are smarter than users should not be paid to persuade them. Agent interconnection Trusted Person Network ": Two users' Instincts can directly negotiate meetings, weekend events, ride sharing Uber, etc., with different permissions set according to their relationship. This is the most imaginative part, and it is also different from the traditional openclaw like assistant part. It is also a key factor in determining whether the service has network effects. Judgment on the Internet He believes that the way most people use software will change in a few years, and the Internet will be rewritten: bookings, travel, subscriptions, etc. will change from "people go to the website" to "say the results, assistants do it". He also acknowledges that computing power is the biggest constraint - about 40% of the time he worries about how computing power can keep up with the daily 10% growth. Regarding competition Admitting that Meta Muse and others are good products, Instinct's approach is to stay within the existing communication channels of users as much as possible, rather than creating a new app. Interview video: https://www. (youtube.com)/watch? v=Am7IWP8IpEc Chinese text version: https://mp.weixin. (qq.com)/s/6njnvKsoUEqFXCwWJA5PRA