Noah，Instinct 的创始人，将他的公司描述为站在“有史以来最激动人心的软件竞赛”的前沿，有望重新定义人们与技术互动的方式。Instinct 这家新公司正在构建一款个人AI助理，它的运作方式不是作为一个新的应用程序，而是一种“全新体验”。它通过现有的通信渠道，如电话、短信和电子邮件，无缝运行，让用户可以像与另一个人交流一样与它互动。这种方法优先考虑社交智能和意识，旨在实现一个看似“简单界面”却蕴含巨大能力的系统。 该产品的影响力已在各种用例中显而易见。用户利用 Instinct 完成各种任务，从衣橱规划和在线购物（生成数千种服装搭配，然后实现一键购买）到个人财务目标（扫描银行账户，端到端地识别并取消不需要的订阅）。一个显著的特点是“Instinct-to-Instinct 网络”，它允许受信任用户的代理直接协调，例如实现无缝的会议日程安排。该网络强调基于信任的不同访问级别，创建了一个动态的互联代理网络。 Noah 强调，Instinct 并非一种“变革性的消费者应用体验”，它不会要求用户信任一种新的愿景。相反，它的运作方式与公众对AI应如何行事的现有预期相符。这种“开箱即用”的特性带来了显著的吸引力，包括通过邀请制系统实现的每日10%的病毒式增长，以及在小用户基数上每年超过10亿美元的交易量。尽管在营销上投入了0美元，却实现了这样的增长，这表明了强大的口碑传播和用户采纳。 信任和用户控制是 Instinct 理念的核心。数据显示，用户需要数周时间才能建立信任，其中 40% 的用户在三周内分享了个人信用卡信息，从而带来了 80% 的留存率。Instinct 坚称用户应始终控制自己的数据并同意共享。为降低风险，公司采用了与代理核心架构解耦的“防火墙”和主动监控系统，以拦截和验证操作，防止恶意互动或幻觉。 其商业模式旨在与用户利益保持一致。Noah 明确表示，Instinct 绝不会为了盈利而影响用户行为，这与广告驱动型平台形成对比。相反，它旨在对用户免费，通过向商家收取统一的交易费率来盈利，类似于支付平台或 Apple Pay、Shopify 等主要电商网站。这种模式利用了 Instinct 不断增长的分销能力及其降低摩擦的能力，可能为合作企业增加交易量。例如，旅游业占 Instinct 当前交易量的 50%，提供了可观的费率分成机会。 构建 Instinct 面临着巨大的挑战，尤其是在计算基础设施方面。每日 10% 的复合增长意味着计算需求每周翻倍，这要求公司需要提前数月采购硬件。Noah 强调了预测和保障如此庞大且迅速增长的计算资源的难度。然而，Instinct 开发了高度优化的推理部署，使得能够以极低的成本服务用户，以远高于标准 API 使用的效率提供前沿智能。 展望未来，Noah 设想更简单的界面，超越基于屏幕的互动，通过耳塞实现语音命令，使技术真正无处不在且主动。他预见到代理将追求更高层次的目标，从个人健身目标到自主运营小型企业。该产品具有自适应性，这意味着它会随着每个用户而演进，学习他们的沟通风格和偏好。最终，Instinct 旨在将分散的软件体验整合到一个单一、直观的界面中，在扩展功能的同时简化互动。

Noah, founder of Instinct, describes his company as being at the forefront of "the most exciting software race ever," poised to redefine how people interact with technology. Instinct, a new company, is building a personal AI assistant that operates not as a new application, but as a "new experience." It functions seamlessly through existing communication channels like phone calls, texts, and emails, allowing users to interact with it as they would with another person. This approach prioritizes social intelligence and awareness, aiming for a "simple interface" that belies its vast capabilities. The product's impact is already evident in diverse use cases. Users leverage Instinct for tasks ranging from wardrobe planning and online shopping (generating thousands of outfits, then enabling one-click purchases) to personal finance goals (scanning bank accounts, identifying and canceling unwanted subscriptions end-to-end). A notable feature is the "Instinct-to-Instinct network," which allows trusted users' agents to coordinate directly, exemplified by frictionless meeting scheduling. This network emphasizes varying levels of access based on trust, creating a dynamic web of interconnected agents. Noah stresses that Instinct isn't a "transformational consumer app experience" where users are told to trust a new vision. Instead, it works in a way that aligns with pre-existing public expectations of how AI should act. This "it just works" factor has driven significant traction, including 10% day-over-day viral growth through an invite-only system, and over $1 billion annually in transaction volume on a small user base. This growth comes despite $0 spent on marketing, indicating strong word-of-mouth adoption. Central to Instinct's philosophy is the concept of trust and user control. Data shows it takes weeks for users to build trust, with 40% sharing a personal credit card within three weeks, leading to an 80% retention rate. Instinct maintains that users should always control their data and consent to sharing. To mitigate risks, the company employs "firewalls" and active monitoring systems, decoupled from the agent's core architecture, to intercept and validate actions, preventing malicious interactions or hallucinations. The business model is designed for user alignment. Noah explicitly states that Instinct will never influence user behavior for profit, contrasting with ad-driven platforms. Instead, it aims to be free for users, monetizing through a blanket transaction take-rate from merchants, similar to payment platforms or major e-commerce sites like Apple Pay or Shopify. This model leverages Instinct's growing distribution power and its ability to reduce friction, potentially increasing transaction volumes for partner businesses. For instance, the travel industry, which accounts for 50% of Instinct's current transaction volume, offers significant take-rate opportunities. Building Instinct presents immense challenges, particularly regarding compute infrastructure. The 10% daily compounding growth means compute demand doubles weekly, requiring the company to procure hardware with months-long lead times. Noah highlights the difficulty in forecasting and securing such massive, rapidly escalating compute resources. However, Instinct has developed highly optimized inference deployments, making it possible to serve users at a very low cost, delivering frontier-level intelligence with significantly more efficiency than standard API usage. Looking ahead, Noah envisions even simpler interfaces, moving beyond screen-based interactions to voice commands via earbuds, making technology truly ubiquitous and proactive. He foresees agents pursuing higher-level objectives, from personal fitness goals to autonomously running small businesses. The product's adaptive nature means it evolves with each user, learning their communication style and preferences. Ultimately, Instinct aims to collapse disparate software experiences into a single, intuitive interface, expanding capability while simplifying interaction.