I reverse-engineered Instinct. Here's the whole machine (twitter.com)

A recent deep dive into Instinct, an AI personal assistant, revealed its impressive capabilities and user-centered design. The author shared their experience of enrolling in dental insurance through Instinct, noting the rapid and efficient interactions—totaling just nine messages—compared to the lengthy processes typically associated with traditional methods. Users have been charmed not only by Instinct's competence but also by its ability to simulate social interactions, making the experience feel more personal and engaging. This combination of functionality with a conversational tone has made it stand out in a crowded market, with some users reporting hundreds of interactions within just days of use. What sets Instinct apart is its blend of operational reliability and interpersonal warmth. Users appreciate its habit of quickly reading messages, providing prompt tapbacks to signal understanding, and its commitment to transparency, informing users when it can't accomplish a request. The AI’s success in building user trust revolves around its ability to follow up, maintain a conversational rhythm, and consistently perform tasks effectively. Internally, it utilizes a system of interconnected technologies that enhance performance while maintaining a human-like persona, marking a significant step forward in AI personal assistants. This focus on user experience could serve as a model for future developments in the AI/ML community.