AdvertiseDonate
AboutRSSJersey CityNew YorkNewarkPhiladelphiaBaltimoreBostonWashingtonRichmond
November 4, 2026
Headlines Buzz · 175 Greenwich Street, 38th floor, New York, NY 10007

Block's new AI model learns language from human voices only

The company says the system runs locally on high-end Macs and aims to better capture the emotional nuance of spoken conversation.

By Harper Hassan ·

Illustration for: Block's new AI model learns language from human voices only
Illustration generated for this article, not a news photograph.

Block, the company led by Jack Dorsey, released an AI model named Dessa that it describes as trained exclusively on human conversation. The announcement highlights a shift away from text-based training data for speech systems.

Ragavan Thurairatnam, who leads the Dessa project, shared sample recordings of the model interacting with people. In one exchange, the model initially mistakes the speaker for someone named Ronaldo before the conversation flows more naturally. The company's website hosts additional examples, including a discussion where the model pivots from Christopher Nolan films to Bryan Singer’s Superman movies.

In another recorded clip, a user describes taking the model on a surprise trip to Paris. The website notes that Dessa responds with what sounds like teary gratitude. A separate clip pairs Dessa with OpenAI’s GPT-Live-1 voice model, where the system sounds more natural than in other samples.

The team states that Dessa is a true speech-to-speech model that derives all its knowledge from human speech. They indicate the system can run locally on a MacBook Pro M4 Max or better hardware. The developers list three motivations for the project: curiosity, improved understanding between humans and AI, and safety.

The team compares the project to how babies learn to talk without seeing written words. They note that while it is difficult to replicate this process in a model, it is possible. The developers argue that current voice assistants flatten speech into text and fail to keep up with human timing. They describe speech as the most natural interface for any intelligence.

The safety argument focuses on the model's ability to detect emotion. The team claims that many AI systems are nearly deaf to the emotional tone in voices. They cite a scenario where a person in serious pain interacts with a voice AI that carries on regardless. The team describes Dessa as a step toward fixing this issue.

Block plans to grant early access to a small group of AI safety researchers. The announcement did not include a timeline for a wider public release.

Source: The Verge

More Tech · Report a business · Got a Story?

Get Headlines Buzz in your inboxWhat’s happening across America. Free, a few times a week, unsubscribe any time.

Get Headlines Buzz in your inboxWhat’s happening across America. Free, a few times a week, unsubscribe any time.