News Summary

AI startup Sesame has unveiled a new Conversational Speech Model (CSM) that has users both amazed and unsettled by its human-like quality. The model, which features voice assistants named ‘Miles’ and ‘Maya’, has been praised for crossing the ‘uncanny valley’ with its expressive and dynamic speech capabilities. Users have reported forming emotional bonds with the AI, reminiscent of scenarios depicted in the film ‘Her’. While the voice model leverages Meta’s Llama architecture to produce near-human speech, it still faces challenges in fully contextual speech generation. Sesame aims to create AI that not only processes requests but engages in genuine dialogue to build trust and confidence over time.