There’s a new voice agent / speech-to-speech benchmark released every other week. So we’ve been looking closely at the design decisions and the assumptions those decisions introduce. On our latest blog: what the model hears, and the choice between human and synthetic speech.

Today we’re announcing Mundo’s $20M Series A, led by @GreatPointVC . Together with our previously unannounced $4M seed, we’ve raised $24M to build the data layer for perceptual intelligence.