![The Power Of Dreams [War & Peas]](https://strongmocha.com/wp-content/uploads/2026/10/the-power-of-dreams-war-peas-featured-200x110.jpg "The Power Of Dreams [War & Peas] 2")
🔍 Read the full analysis: Block’s Dessa’s Voice Sounds Human. Should That Change What We Believe? on ThorstenMeyerAI.com
Get audio and creator gear 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 can listen and speak at once, recognize emotion and run locally on powerful consumer hardware. The demonstrations are Block-selected recordings; no public model, technical paper or independent test is available. Its humanlike voice may improve conversation, but it also sharpens concerns about impersonation and voice-based fraud.
Block announced Dessa, a speech-to-speech research model that the company says can respond with humanlike timing, laughter and emotional cues. The demonstrations have drawn attention to both its potential for more natural voice interactions and the risk that realistic AI voices could make impersonation scams harder to detect, but the model has not been independently tested and is not publicly available.
Block, the company behind Square and Cash App, published Dessa’s project page on October 23. The company says the model was trained from scratch on recorded human conversation, without text or pretrained models. It describes Dessa as full-duplex: able to listen while speaking, interrupt, and use conversational cues such as backchannels. Block also says it can handle emotion, sarcasm and laughter, and that inference runs faster than real time on a MacBook Pro M4 Max or better. It has not provided a latency figure.
The public evidence so far consists of recordings selected by Block. In them, Dessa laughs during a sentence, responds to a surprise and helps a speaker recall a film title. Block says the clips are live recordings and that nobody steered the model. The company has offered a waitlist and early access to a small group of AI safety researchers, but has announced no public weights or API.
Outside coverage has been more guarded than the launch claims. The Verge described some exchanges as occasionally awkward and uncanny, while saying Dessa came across as more natural than another voice model in one comparison. TechCrunch reported that it was unclear whether emotion in the audio was being guided by tags. CoinDesk noted that Block had not disclosed the model’s size, benchmark results or a public release date. Those observations are based on listening to demonstrations, not independent access to the system.
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
Humanlike Voices and Fraud Checks
A more fluid voice interface could make automated services feel easier to use, particularly when callers need to describe an urgent or emotional situation. Block argues that current voice AI can miss emotional cues and continue speaking when a person is in pain. Dessa’s conversational timing and expressive audio could address that limitation, if the company’s claims hold up in broader testing.
The same qualities could also make a synthetic caller more persuasive. The FBI’s 2025 Internet Crime Report, published in April 2026, recorded 22,364 complaints mentioning AI and $893 million in reported losses. The FBI said victims lost more than $5 million to distress scams using cloned voices to imitate loved ones in trouble. These figures describe reports and losses in 2025; they do not establish that Dessa was used in any fraud.
For readers, the practical lesson is not that every realistic voice is fraudulent, but that a familiar-sounding voice is not proof of identity. Families can agree on a private code word and call back using a number they already trust. Businesses can require a separate verification step for payment requests rather than treating a voice alone as authorization. Such precautions address existing voice-cloning risks, regardless of whether Dessa is released.
Products Worth Considering
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
As an affiliate, we earn on qualifying purchases.
What Block Has Demonstrated So Far
Block’s announcement describes a research project from a small AI lab at the company. The team is associated with Dessa, a Toronto deep-learning startup acquired by Square in 2020, according to CoinDesk. Block says the new model learned from hundreds of times fewer words than large language models require and was trained using one NVIDIA GB200 NVL72. Those are company-provided descriptions; the announcement does not include a paper or benchmark data that would allow outsiders to assess them.
Co-founder Jack Dorsey said that talking with Dessa feels like talking with a person, citing emotion, nuance, pauses and natural cadence. Project lead Ragavan Thurairatnam called it “a GPT3 moment for speech” and said it could run on consumer GPUs. The claims help explain the excitement around the recordings, but they do not establish intelligence, emotional understanding or general reliability. A voice that sounds expressive may still lack human feelings or judgment.
Block’s own project page calls efficient learning of new languages an open question. The supplied material also does not identify a supported language list or describe a public testing process. In EU policy, Article 50 of the AI Act is described in the source material as applying from August 2, 2026, with transparency requirements relevant to certain AI-generated or manipulated content. The exact obligations depend on the use and circumstances; the announcement does not explain how Block plans to address disclosure or labeling for Dessa.
“Talking with it feels like talking with a person…full of emotion, nuance, pauses, and a natural cadence.”
— Jack Dorsey, Block co-founder
Independent Tests Still Missing
As of October 26, no technical paper, model card, benchmark results, license or public release date had been identified in the source material. No outside researcher or reporter has been able to run the model. That leaves key claims about performance, language coverage, reliability and hardware requirements unverified.
It is also unclear how the system handles consent, voice imitation, watermarking or disclosure to people speaking with it. Block says Dessa chooses its own voice and personality, but that does not answer whether it can imitate a particular person or what safeguards would prevent that. The source material does not establish that the model has been used in fraud, nor does it show that its demos reflect performance across unscripted conversations.
Release Details and Safety Evidence
The next useful evidence would be independent access and technical documentation: a model card, benchmarks, clear hardware and language specifications, and a description of safeguards. Block has said early access is available to a small group of AI safety researchers, but the company has not announced when wider testing or a public release will take place.
Until those details emerge, readers and organizations should treat the demonstrations as evidence of what Block chose to show, not proof of broad capability. Companies can verify sensitive requests through a known callback number or another established procedure. Families can agree on a code word and avoid sending money in response to an urgent call without checking through a separate channel. Those steps remain useful even if Dessa never becomes a consumer product, because voice cloning and impersonation scams already exist.
Key Questions
What is Block’s Dessa?
Dessa is a speech-to-speech research model announced by Block. The company says it can listen and speak at the same time and respond with conversational cues such as laughter and pauses.
Can the public use or test Dessa now?
Not broadly. As of October 26, Block had announced a waitlist and early access for a small group of AI safety researchers, but no public weights or API. Outsiders have not independently run the model.
Has Dessa been shown to commit fraud?
The source material provides no evidence that Dessa has been used in fraud. The concern is that human-sounding synthetic voices could make impersonation more convincing; voice-cloning scams were already reported before this announcement.
How can I check whether an urgent call is genuine?
Hang up and call the person back using a number you already know. Families can also agree on a private code word. For businesses, verify payment or account changes through a separate, established channel rather than relying on voice recognition alone.
Source: ThorstenMeyerAI.com
