Observed Signal · Oct 24, 2026 · Product Launch · Source: techcrunch · Impact: 3/5 · Sentiment: Positive
Block Unveils Human-Sounding Voice Model Dessa
Block, the fintech company led by Jack Dorsey, has released a new voice model called Dessa from its research lab. The model is trained fully on human speech and can run locally on a consumer-grade MacBook M4 Max. Dessa is a full-duplex model, enabling simultaneous listening and speaking, similar to OpenAI's GPT-Live. Unlike many voice models that rely on text pipelines and voice tags, Dessa embeds emotion directly into the model, as demonstrated in samples where the bot conveys tiredness or emotional distress. The company claims this leads to more natural interactions, addressing the loss of emotional information in text-based approaches. Currently, Dessa is not publicly accessible, and Block has invited safety researchers to test the model for potential risks, such as deceptive voice agents. This development could enable more cost-effective voice experiences for businesses, building on Square's existing voice AI for ordering.
The release of a full-duplex, locally running voice model from a major fintech player signals advancement in voice AI applications, potentially lowering cost and increasing adoption in advertising and customer service contexts.
Track Block Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your Polaris7 workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Block released a new voice model called Dessa.
- The model runs locally on a MacBook M4 Max.
- Dessa is a full-duplex model that can talk and listen simultaneously.
- The model embeds emotion, trained fully on human speech.
- There is no public access yet; safety researchers are being recruited.
Connected Companies & Entities
3 Entities mapped“Block released a new voice model called Dessa out of the company’s research lab.”
“Dessa is a full-duplex model, which means it can talk and listen at the same time, just like the GPT-Live model OpenAI released earlier this...”
“A lot of models like ElevenLabs v4 or Fish Audio Drama 3 use voice tags to drive emotion in speech generation.”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
ChatGPT's GPT Live Voice Sounds Human, Causing Unease
OpenAI has updated ChatGPT's GPT Live audio so the assistant inserts human-like filler words and hesitations (e.g., 'hmm', 'ja', 'okay') to make spoken conversations sound more natural. The author reports an uncanny-valley unease: the convincingly human voice can encourage misplaced trust and appears to be a deliberate design choice to increase engagement. The piece cites AI researcher Katharina Zweig on the lack of an innate human 'sensor' for detecting non‑human intelligence, amplifying concerns about trust. While the change lowers friction and can make voice interactions more productive, the article warns users to avoid sending highly personal or sensitive information to cloud‑based conversational AI. The author emphasizes the trade-off between productivity gains and ethical, privacy and trust risks. The update builds on earlier Advanced Audio Mode improvements and the article was published 2026-08-16.
OpenAI Launches GPT-Live-1 Voice Model in API
OpenAI has launched GPT-Live-1 in its API, a voice-native model designed for building voice-enabled applications and business workflows. The model handles listening and speaking simultaneously, reducing latency by eliminating the need for chained speech-to-text, reasoning, and text-to-speech components. It features improved interruption handling, background noise resilience, and the ability to delegate reasoning and tool calls to backend models like GPT-6 Astra or third-party models. The API includes telephony support for full-duplex voice agents. Pricing is set at $0.05 per minute for the voice layer, with backend models billed separately. Early evaluations show an 80% reduction in interruptions for language tutoring compared to turn-based systems.
OpenAI launches GPT‑Live voice models
OpenAI introduced GPT‑Live, a new full‑duplex family of voice models powering ChatGPT Voice that can listen and speak simultaneously and delegate complex tasks to other models in the background. GPT‑Live (including GPT‑Live‑1 and GPT‑Live‑1 mini) adds live translation, visual answer cards and improved noise suppression; the mini version will be available to free ChatGPT users. OpenAI says GPT‑Live can hand off deeper analyses to larger models such as GPT‑5.5 (and an upcoming GPT‑5.6) and then return results into the ongoing conversation. The models are rolling out on ChatGPT for web, iOS and Android, with API access planned and developer registration open. The article references hands‑on testing by AI expert Jens Polomski and notes more than 150 million people already use ChatGPT’s voice/dictation features.
Receive all future signals for Block & competitors in your briefing
Never miss material M&A deals, leadership changes, account wins, or product rollouts. Polaris7 monitors 20,000+ primary sources 24/7, delivering verified telemetry directly into your executive briefing.