Live TVMarketsFinanceEconomicsIndustriesTechReal EstateSmall BusinessPoliticsVideo
Top SecuritiesDOW (DIA) $521.37 ▼1.24%S&P 500 (SPY) $759.97 ▼0.52%NASDAQ 25,955 ▼0.21%10Y 5.00% ▲4 bpBITCOIN $80,256 ▼1.15%NASDAQ 100 (QQQ) $704.40 ▼0.01%30Y MORTGAGE 7.32% ▼0.08
Saturday, October 24, 2026 Edition · Raleigh 64° Mostly Cloudy Watch Live Sign in Subscribe
GENIUS EYEGET NewsA service of The GET Network, Inc.
Markets
Tech

Block releases Dessa, a human-like voice model running locally on MacBook M4 Max

Block's new voice AI model Dessa can talk and listen simultaneously with natural emotion and nuance.

Oct 24, 2026, 12:19 PM By the Genius Eye newsroom (Tech Reporter) · Fact-checkedHow we report
Photo: Coolcaesar / CC BY-SA 3.0, via Wikimedia Commons

Block, led by Jack Dorsey, introduced a new voice model named Dessa that runs locally on devices like the MacBook M4 Max, according to TechCrunch.

Dessa is a full-duplex speech model, meaning it can talk and listen at the same time, similar to OpenAI's GPT-Live model.

Unlike many current speech models that rely on text conversion and voice tags to simulate emotion, Dessa is trained fully on human speech and integrates emotion directly into its responses.

Block's researchers demonstrated Dessa's ability to express different emotions and natural speech patterns, such as sounding tired when stating it was tired from work.

The company emphasized that current voice AI often loses emotional context by converting voice to text, which they believe worsens user experience and can be dangerous.

Block aims to provide a more authentic and emotionally nuanced voice AI interaction with Dessa, which is powered by a single NVIDIA GPU and conversational data.

Sources
  1. TechCrunch: Block’s new voice model sounds eerily human
We read public reporting and official data, confirm what we can, and write our own story. See something wrong? Request a correction.

Comments

No comments yet.

Sign in to comment