It's time for new training regimes in speech. We believe audio-native RL environments with fine-grained rubrics (naturalness x task completion) is part of the answer. We've been doing research on this and can soon release a preview of our own post-training. Truly exciting times!
We believe most interaction with AI will happen through speech. But models still talk like machines.
This is a data problem.
Today we're launching @agenticdataco (YC F26): real-world speech data + audio-native RL environments so speech models sound and act human.
Our OpenYap Turn model takes #1 for end-of-turn on @sesame's TurnBench
Trained on 1,984rs of real conversation between friends and family.
Read the writeup: https://t.co/6W1qM0BkfB
today, The Agentic Data Company is releasing the world’s largest free conversational dataset.
- 1,000 hours English dual-channel
- natural conversations between friends/family
free for commercial and research use. https://t.co/ON2PNIF3wd