🔍 Read the full analysis: Block’s Dessa Laughs, Hesitates And Interrupts Like A Person. Why A Human-Sounding Voice Is No Longer Proof Of Anything on ThorstenMeyerAI.com
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Block announced Dessa, a speech-to-speech research model that the company says was trained on recorded human conversation and can handle emotion and interruptions. Its demos have not been independently verified, and Block has not published a paper, benchmarks or a public release date. The lifelike voice raises questions about accessibility and fraud, but the model’s capabilities and safeguards remain untested outside the company.
Block announced Dessa, a research model designed to listen and speak in real time, with company demos showing it laughing, reacting emotionally and interrupting naturally. The demonstrations offer a glimpse of more human-sounding voice AI, but the model has not been independently tested, and Block has not published its technical details or made it publicly available.
Block, the company behind Square and Cash App, published a project page for Dessa on October 23. The company says it trained the model from scratch using only recorded human conversation, with no text data or pretrained models. Block describes Dessa as a speech-to-speech system that listens while it talks, rather than converting speech into text before generating a reply.
In recordings selected and published by Block, Dessa laughs mid-sentence, responds to a surprise and helps a speaker recall a film title. Block says the clips are live conversations and that no one steered the model during them. The company also says it can run faster than real time on a MacBook Pro M4 Max or better, but has not provided a latency figure. Its team says the model was trained using one NVIDIA GB200 NVL72; Block has not released the model’s size or benchmark results.
The demos are not a substitute for outside testing. The Verge described some exchanges as occasionally awkward and uncanny, while saying Dessa seemed more natural than another model used for comparison. TechCrunch said it was unclear whether emotion in the output was guided by tags. As of October 26, Block had announced a waitlist and early access for a small group of AI safety researchers, but no public release date, API or model weights.
Infographic 01 · Dessa fact sheet
What Block claims, what was heard, what is missing.
Every capability below comes from Block or its selected recordings. Nobody outside Block has run the model yet.
| Topic | What we know | Status |
|---|---|---|
| Training data | Human conversation only: “no text, no pretrained models” | Block’s claim |
| Conversation | Full-duplex: listens while it talks, backchannels, cuts in | Block’s claim |
| Emotion | Hears and speaks laughter, sarcasm, sighs; unclear if steered by tags (TechCrunch) | Block’s claim |
| Speed | “Faster than real time”; no latency figure | Block’s claim |
| Hardware | Trained on one NVIDIA GB200 NVL72; runs on a MacBook Pro M4 Max or better | Block’s claim |
| Demos | Some chats “a little awkward and uncanny”, more natural than GPT-Live-1 (The Verge) | reporters heard it |
| Size, benchmarks, release date | Not disclosed (CoinDesk) | not disclosed |
| Paper, model card, license, weights | None found as of 26 Oct 2026 | not disclosed |
| Languages | Not listed; Block calls learning new languages an open question | not disclosed |
| Voice cloning safeguards | No statement on watermarking, consent checks or imitating a specific person | not disclosed |
Who can use it
research preview
- Public: a waitlist; “launch links are coming soon”
- Early access: “a small group of AI safety researchers”, institutional email required
- Business use: none announced; Square already uses voice AI for ordering (TechCrunch)
Who built it
from Block’s page
- A “(very) small AI research lab at Block” reporting to Jack Dorsey
- Founded by former members of Dessa, the Toronto startup Square acquired in 2020 (CoinDesk)
- Lead Ragavan Thurairatnam: “a GPT3 moment for speech”
Sources: Block, Dessa project page (23 Oct 2026); Ragavan Thurairatnam on X (23 Oct 2026); The Verge (23 Oct 2026); TechCrunch (24 Oct 2026); CoinDesk (24 Oct 2026).
Powered by Thorsten Meyer AI
Human-Sounding Voices and Fraud Risk
A voice that handles pauses, laughter and interruptions may make automated services feel more responsive. The same traits could make an impersonation harder to spot by ear. That risk does not depend on Dessa being released: voice cloning is already part of reported fraud, and a more natural conversational style could remove one cue some people rely on when judging a call.
The FBI’s 2025 Internet Crime Report, published in April 2026, recorded 22,364 complaints mentioning AI and $893 million in reported losses. The report also counted more than $5 million in losses from distress scams involving cloned voices of loved ones. Those figures describe reported complaints and losses, not the prevalence of Dessa or proof that its model has been used in a crime.
For businesses and families, the practical implication is that a familiar-sounding voice should not be treated as proof of identity. Organizations can require callbacks to known numbers before approving payments, while families can agree on a private code word for emergencies. Such steps address voice impersonation generally; Block has not said that Dessa can imitate a particular person.
USB microphone for AI voice testing and speech recording
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What Block Says Dessa Can Do
Block presents Dessa as a departure from speech systems that rely on a text transcript. The company says those systems can discard cues such as sarcasm, laughter or a frustrated sigh. It describes Dessa as able to hear and express those cues while managing turn-taking in a conversation. These are Block’s descriptions of the model, not results from a published independent evaluation.
Jack Dorsey, Block’s co-founder, praised its conversational style after the announcement. Dessa lead Ragavan Thurairatnam called the work “a GPT-3 moment for speech” and said it could run on consumer GPUs. The project team is a small research lab at Block, formed by former members of the Toronto deep-learning startup Dessa, which Square acquired in 2020, according to CoinDesk.
Coverage of the demos has been mixed rather than uniformly impressed. The Decoder cautioned that sounding emotional is not the same as being emotional or intelligent. Block itself has framed the project as research, and the company’s page leaves open questions including how efficiently the model can learn other languages.
“Talking with it feels like talking with a person…full of emotion, nuance, pauses, and a natural cadence.”
— Jack Dorsey, Block co-founder
closed-back headphones for speech and voice evaluation
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Testing, Access and Safeguards
Key facts about Dessa remain unavailable. Block has not published a technical paper, model card, license or benchmark results, and outsiders have not been given broad access to test the system. The company has not listed supported languages or announced when the model might be released publicly.
It is also unclear whether the model can imitate a specific person, what consent checks it uses, or whether its output is watermarked. Block says Dessa selects its own voice and personality, but that does not answer whether other forms of voice imitation are technically possible or restricted. The existing demos cannot establish how the system performs across different speakers, noisy calls or adversarial prompts.
MacBook Pro M4 Max for local AI model development
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Independent Access Is the Next Test
The next meaningful milestone is access for researchers who can test the model outside Block’s demonstrations. The company has said a small group of AI safety researchers will receive early access, but it has not identified them or provided a schedule for wider testing. Their findings could help clarify performance, limits and the risk of misuse.
Until Block publishes technical documentation or expands access, claims about Dessa’s efficiency, emotional abilities and readiness for practical use should be treated as company statements. The project page and any future disclosures may also answer questions about languages, consent measures, watermarking and release plans. For now, the confirmed development is narrower: Block has shown recordings of a promising research system, but its public capabilities and safeguards remain unverified.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is Block’s Dessa?
Dessa is a speech-to-speech research model that Block says was trained from scratch on human conversation. The company says it can listen and speak at the same time and respond to vocal cues such as laughter and sarcasm.
Can the public use Dessa now?
No public release has been announced. As of October 26, Block had a waitlist and said a small group of AI safety researchers would receive early access; it had not announced an API or published model weights.
Has an independent group verified the demos?
Not broadly. Reporters have listened to Block’s selected recordings, but the model has not been made available for outside testing at scale, and Block has not published benchmark results.
Does Dessa clone a specific person’s voice?
That has not been established. Block says the model chooses its own voice and personality, but has not detailed whether it can imitate specific people or what safeguards would prevent that.
How can people reduce the risk of voice scams?
Do not rely on a familiar voice alone to verify identity. For urgent requests involving money or personal information, hang up and call a number already known to you; families may also agree on a private code word.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
