Ex-Google X trio wants their AI to be your second brain — and they just raised $6M to make it happen

A trio of former Google X scientists is on a mission to create an advanced AI application that serves as a virtual second brain, designed not for science fiction but for practical daily use. Their startup, TwinMind, has successfully raised $5.7 million in seed funding and launched versions of their app for both Android and iPhone. This innovative tool functions by passively listening to conversations in the background, with user consent, to develop a personal knowledge graph. Founded in March 2024 by Daniel George, along with his ex-Google X colleagues Sunny Tang and Mahi Karim, TwinMind captures ambient audio to transform spoken thoughts, meetings, and discussions into structured memories. The app generates AI-enhanced notes, task lists, and responses by processing audio in real-time, all while operating offline. Remarkably, it can continuously record for up to 17 hours without significantly draining the device's battery. In addition to its core functionalities, TwinMind allows users to back up their data, ensuring that vital conversations can be restored if a device is lost, although users have the option to opt-out. The app also boasts real-time translation capabilities in over 100 languages, setting it apart from other AI note-taking applications. What differentiates TwinMind from competitors like Otter and Fireflies is its ability to capture audio throughout the day without requiring constant user engagement. The team developed a native service in pure Swift for iPhone, which avoids the limitations of cloud-based processing that many competitors rely on. George, who left Google X in 2020, conceptualized TwinMind while working at JPMorgan as the Vice President and Applied AI Lead. His experience attending numerous meetings daily led him to create a script that captured and transcribed audio, which then interacted with ChatGPT to produce actionable insights. After sharing his creation, he received feedback indicating a strong desire for a personal device solution that operates discreetly during meetings. The startup also offers a Chrome extension that enhances context gathering by analyzing browser activity and interpreting information across platforms like email and Slack. TwinMind utilized this tool to aid in selecting interns from over 850 applicants this summer, showcasing its effectiveness. Currently, TwinMind has more than 30,000 users, with a growing number from the U.S., India, Brazil, the Philippines, and several other regions. The user base includes professionals, students, and individuals using the app for personal projects. Emphasizing user privacy, George clarified that TwinMind does not train its models on personal data or store audio recordings; instead, it deletes audio immediately while retaining only transcribed text locally. With a team of 11, TwinMind plans to enhance user experience by hiring designers and establishing a business development team to expand its API offerings. The recent funding round involved significant investments from Streamlined Ventures, Sequoia Capital, and other notable investors, valuing the startup at $60 million post-money. In addition to its apps, TwinMind has introduced a new AI model, the Ear-3, which supports over 140 languages and features improved accuracy in speaker recognition. The application will soon be available through an API for developers and enterprises, while the existing free version continues to offer unlimited transcriptions and on-device speech recognition. TwinMind has ambitious plans for growth and user acquisition in the near future.