Block unveils Dessa voice model trained only on speech

A voice model trained only on speech data tests whether conversation audio alone improves naturalness.

Jack Dorsey's Block released Dessa, a voice AI model trained solely on human conversations rather than text, aimed at more natural-sounding speech.

Name

Dessa

Domain

creative

Read it at The Verge →

Coverage: 13 sources, in order

1. 23 Oct 19:32 The Verge Read the article →
2. 23 Oct 20:27 UA.NEWS Read the article →
3. 23 Oct 20:51 FourWeekMBA Read the article →
4. 23 Oct 21:03 news.lavx.hu Read the article →
5. 23 Oct 23:36 Reddit Read the article →
6. 23 Oct 23:46 RuntimeWire Read the article →
7. 24 Oct 05:43 The Indian Express Read the article →
8. 24 Oct 05:49 Awaz The Voice Read the article →
9. 24 Oct 06:23 TheStreet Read the article →
10. 24 Oct 06:30 BW Businessworld Read the article →
11. 24 Oct 06:58 The Economic Times Read the article →
12. 24 Oct 07:05 India Today Read the article →
13. 24 Oct 07:42 Exchange4Media Read the article →

← Front page