Reaction
Bland introduces Tofu-1, a voice model designed to follow the caller—not the TV
Bland says its speech-to-speech model handles interruptions and background voices; research preview access is limited.
TLDR
With Tofu-1, Bland is pitching phone AI that listens while speaking and distinguishes callers from background voices. The company is accepting research-preview requests from a limited number of enterprises, targeting banking, healthcare, insurance and government. Its example of ignoring a TV’s answer to a policy-cancellation question illustrates intended behavior—it is not a recorded model output.
Combined views
—
2 Sources, first seen ago
Comments