This a16z AI podcast episode interviews Ankit Kumar, Sesame's co-founder and CTO, about the Sesame Research Preview, a new approach to conversational AI. The conversation begins with Ankit's initial reaction to the positive reception of the preview, then delves into the development process, focusing on the challenges of balancing quantitative ML metrics with qualitative user experience. Key discussion points include the technical decisions behind Sesame's design (e.g., moving towards transcription-free audio processing), its competitive advantages, and the future roadmap, which includes open-sourcing components of the speech generation model and developing a full-duplex conversational AI for AR glasses. Ankit emphasizes the importance of focusing on creating a delightful user experience, rather than solely prioritizing raw technological capabilities. The podcast concludes with a discussion of Sesame's future plans, including app development and the potential for third-party developer integration.
