What Laughter And Hesitation Mean In Block’s Dessa’s AI Voice
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: What Laughter And Hesitation Mean In Block’s Dessa’s AI Voice on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

Block announced Dessa, a speech-to-speech research model whose company-selected demos feature laughter, pauses and conversational interruptions. The model has not been independently tested or released publicly, so its capabilities and risks remain unverified.

Block has published demos of Dessa, a speech-to-speech research model that laughs, pauses and responds to interruptions in conversation. The recordings suggest a more natural-feeling voice interface, but the model has not been independently tested, and Block has not announced a public release, leaving its performance and safeguards unverified.

Block announced the project on October 23. The company says Dessa was trained from scratch on recorded human conversation, without text or pretrained models, and can listen while speaking. Block describes it as recognizing and producing cues such as laughter, sarcasm and pauses, which can be lost when speech is converted into text before a system responds.

The company’s recordings include Dessa laughing during a sentence, helping a speaker recall a film title and responding to an anniversary surprise. Block says each clip is a live conversation and was not steered by a person. Co-founder Jack Dorsey wrote that talking with it feels like talking with a person, citing its emotion, nuance, pauses and cadence. These are Block’s claims and selected examples, not results from an independent evaluation.

Block also says the model runs faster than real time on a MacBook Pro M4 Max or better, but it has not given a latency figure. It says training used one NVIDIA GB200 NVL72. As of October 26, there is no public paper, model card, benchmark, license or API; access is limited to a waitlist and a small group of AI safety researchers. The Verge described some conversations as occasionally awkward and uncanny, while TechCrunch said it was unclear whether emotion was being shaped by tags.

At a glance
reportWhen: Announced October 23, 2026; public acce…
The developmentBlock published demonstrations of Dessa, a voice AI research model, on October 23, 2026.
Crypto market snapshot
Fear & Greed Index
50/100 — Neutral
Bitcoin BTC$80,153▼ 0.4%
Ethereum ETH$2,324▼ 1.0%
Tether USDT$0.9993▲ 0.0%
BNB BNB$741.44▼ 0.3%
XRP XRP$1.29▼ 0.7%
USDC USDC$0.9998▲ 0.0%
Solana SOL$101.36▼ 0.4%
TRON TRX$0.3299▲ 0.4%
Live data · CoinGecko · alternative.me (24h change)
Thorsten Meyer AI
Status · 26 Oct 2026

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.

1NVIDIA GB200 NVL72 rack used for training, per Block
Notext in the training data, per Block
0independent tests published so far
M4 MaxMacBook Pro class Block says runs it locally
TopicWhat we knowStatus
Training dataHuman conversation only: “no text, no pretrained models”Block’s claim
ConversationFull-duplex: listens while it talks, backchannels, cuts inBlock’s claim
EmotionHears and speaks laughter, sarcasm, sighs; unclear if steered by tags (TechCrunch)Block’s claim
Speed“Faster than real time”; no latency figureBlock’s claim
HardwareTrained on one NVIDIA GB200 NVL72; runs on a MacBook Pro M4 Max or betterBlock’s claim
DemosSome chats “a little awkward and uncanny”, more natural than GPT-Live-1 (The Verge)reporters heard it
Size, benchmarks, release dateNot disclosed (CoinDesk)not disclosed
Paper, model card, license, weightsNone found as of 26 Oct 2026not disclosed
LanguagesNot listed; Block calls learning new languages an open questionnot disclosed
Voice cloning safeguardsNo statement on watermarking, consent checks or imitating a specific personnot 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

Natural Voices and Fraud Risks

More fluid speech could make voice assistants easier to talk with, particularly when users interrupt, hesitate or express distress. Block argues that voice systems often miss emotional cues and may continue responding inappropriately when someone is in pain. The same humanlike cues, however, could make a synthetic voice harder to distinguish from a real caller.

That risk is relevant because voice impersonation scams already exist; they do not depend on Dessa. The supplied source cites FBI figures for 2025: 22,364 complaints mentioning AI and $893 million in reported losses. It also reports more than $5 million in losses from distress scams involving cloned voices of loved ones. Those figures describe reported complaints and losses, not a measure of Dessa’s capabilities or its role in any fraud.

The practical point is that a familiar-sounding voice should not be treated as proof of identity. Organizations handling payments or sensitive information may need to rely on verification beyond voice recognition. For families, a separate code word and calling back using a number already saved can help check an urgent request. Those steps address existing impersonation risks; the source does not establish that Dessa has been used in a scam.

Products Worth Considering

Apple MacBook Pro 2026 14.2 Laptop with M5 Pro chip, 15‑core CPU, 16‑core GPU, 24 GB Unified Memory, 1 TB SSD, Wi‑Fi 7 – Silver

Apple MacBook Pro 2026 14.2 Laptop with M5 Pro chip, 15‑core CPU, 16‑core GPU, 24 GB Unified Memory, 1 TB SSD, Wi‑Fi 7 – Silver

Apple MacBook Air 13.6-inch Laptop with M5 Chip, 24GB Memory, 1TB SSD, Midnight

Apple MacBook Air 13.6-inch Laptop with M5 Chip, 24GB Memory, 1TB SSD, Midnight

As an affiliate, we earn on qualifying purchases.

How Dessa Reached Its Demo Stage

Dessa is a project from a small research group at Block, the company behind Square and Cash App. The group was founded by former members of Dessa, a Toronto deep-learning startup acquired by Square in 2020, according to CoinDesk. The shared name refers to the earlier startup as well as the newer research project.

Block presents the model as a speech-first system: it says Dessa learned from conversation audio rather than text and can handle speaking and listening at the same time. The company has not published enough technical detail for outsiders to verify those claims, compare results with other systems or assess how reliably it handles different speakers and situations.

The initial reactions have focused on both usefulness and impersonation. Block frames emotional awareness as a way for voice systems to respond better to people. Critics, including filmmaker Justine Bateman, have questioned why an AI voice needs to sound human. The demos make the tension visible, but they do not settle it: a lifelike performance is not evidence that a system understands emotion or has human feelings.

“Talking with it feels like talking with a person… full of emotion, nuance, pauses, and a natural cadence.”

— Jack Dorsey, Block co-founder

What Testing Has Not Established

Outside Block has not been given general access to test Dessa, and there are no published benchmarks or technical paper in the supplied material. It is not clear how the system performs across languages, accents, noisy calls or extended conversations, or how often its interruptions and emotional cues are appropriate. Block has itself identified efficient learning of new languages as an open question.

Other unresolved points include whether emotional delivery is controlled by tags, whether the model can imitate a specific person, and what safeguards or watermarking it uses. Block says Dessa selects its own voice and personality, but that does not answer whether users could prompt it to resemble someone else. The source says there is no announced release date, and it does not establish that the system has been used to deceive anyone.

Release Details and Independent Review

The next clear milestones would be publication of technical documentation, independent testing and details on access, safeguards and licensing. Block has not announced when those might arrive. Until then, claims about speed, training efficiency and conversational realism should be treated as company-reported capabilities, not independently verified results.

As voice AI develops, users and organizations can continue to verify unexpected requests through a separate channel rather than relying on a caller’s voice alone. The source also points to disclosure rules under Article 50 of the EU AI Act, which it says apply from August 2, 2026; how any particular service must comply depends on its use. No release plan or policy change for Dessa has been confirmed.

Key Questions

What is Block’s Dessa?

Dessa is a speech-to-speech research model that Block says was trained on recorded human conversation and can listen while it speaks. The company has shown demos but has not made the model generally available.

Has Dessa been independently tested?

Not publicly, according to the source. As of October 26, 2026, Block had published no paper, model card or benchmark, and outside researchers had not run the model.

Can Dessa clone a particular person’s voice?

The available material does not establish that. Block says Dessa chooses its own voice and personality, but has not provided a clear public account of whether it can imitate a specific person or what protections apply.

Does Dessa create a new voice-scam threat?

The demos do not show that Dessa has been used in fraud. Voice-cloning scams already exist, so a familiar-sounding caller should not be treated as identity verification. Check urgent requests through a known, separate contact method.

Source: ThorstenMeyerAI.com

Nothing in this article is financial or investment advice. Cryptocurrency and precious-metal investments carry significant risk — do your own research and consider a licensed advisor.
You May Also Like

The Future Of AI Industry Disruption: Insights From ByteDance’s Strategy

ByteDance Seed describes its AI development approach as ‘slow first, fast afterwards,’ signaling a strategic shift that could impact industry competition.

13 Best Portable SSDs For Compact, Convenient Storage In 2027

A 13-drive portable SSD guide compares Samsung and SanDisk models by capacity, advertised speed, USB compatibility and intended use.

AI Workflow Automation In 2027: 15 Tools To Know

A roundup of 15 books on AI workflow automation, from Microsoft Power Automate and no-code guides to accounting, marketing and NetOps.

Discover The 10 Best Mirrorless Cameras With AI For 2026

Discover the 10 best AI-enabled mirrorless cameras for 2026, featuring models suited for beginners, professionals, and content creators. Updated list with detailed insights.