{"id":184,"date":"2026-11-03T11:47:59","date_gmt":"2026-11-03T11:47:59","guid":{"rendered":"https:\/\/technobuzzsystems.com\/blog\/?p=184"},"modified":"2026-11-03T12:44:32","modified_gmt":"2026-11-03T12:44:32","slug":"dessa-speech-to-speech-ai","status":"publish","type":"post","link":"https:\/\/technobuzzsystems.com\/blog\/dessa-speech-to-speech-ai\/","title":{"rendered":"Dessa AI: Block\u2019s Breakthrough Full-DuplexSpeech-to-Speech Mode"},"content":{"rendered":"\n<figure class=\"wp-block-image alignwide size-large wp-duotone-unset-1\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"683\" src=\"https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-5-1024x683.png\" alt=\"Dessa AI: Block\u2019s Breakthrough Full-DuplexSpeech-to-Speech Mode\" class=\"wp-image-196\" srcset=\"https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-5-1024x683.png 1024w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-5-300x200.png 300w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-5-768x512.png 768w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-5.png 1536w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">&#8220;The launch of Dessa AI represents an interesting development in the evolution of speech-to-speech artificial intelligence.&#8221; Dessa speech-to-speech AI is Block\u2019s full-duplex voice model, designed to process and generate speech directly from audio. In this article, we explore its reported architecture, potential applications, limitations, and how it compares with AI voice tools such as ChatGPT Voice. Voice AI traditionally works by converting speech to text, running a language model, then synthesizing speech again \u2013 a process that strips away tone, emotion, and spontaneity. Block\u2019s new Dessa AI flips this pipeline on its head. Announced in October 2026, Dessa is a full-duplex, end-to-end speech-to-speech model: it takes raw audio in and generates audio out, with no text step at all. In practice, this means Dessa can listen and speak simultaneously (like a human), laugh mid-sentence, backchannel (e.g., \u201cyeah,\u201d \u201cuh-huh\u201d), and adjust volume on cue. The system was trained <em>exclusively<\/em> on recorded human conversations \u2013 no text transcripts, no pretrained models \u2013 and runs on a single high-end GPU (NVIDIA GB200 NVL72).<\/p>\n\n\n\n<figure class=\"wp-block-image alignwide size-large wp-duotone-unset-2\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"576\" src=\"https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-3-1024x576.png\" alt=\"Microphone representing voice input for AI conversations with audio waveforms and AI symbols.\" class=\"wp-image-190\" srcset=\"https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-3-1024x576.png 1024w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-3-300x169.png 300w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-3-768x432.png 768w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-3-1536x864.png 1536w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-3.png 1672w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">Block\u2019s lab reports that <em>every<\/em> behavior of Dessa was learned from speech alone. In demos, when a human partner hesitates, Dessa can finish their sentence correctly; when something is funny, it laughs mid-utterance. Because it never converts voice to text, it naturally captures nuances like sarcasm, sighs, tears, and emotion that text-based systems miss. Block positions Dessa as \u201cthe most authentic full-duplex speech model ever built,\u201d suggesting it could make voice interactions far more natural.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">However, it\u2019s important to note that Dessa is currently a <strong>research preview<\/strong>, not a product. Block has released demo recordings and invited a small group of AI safety researchers to experiment with it \u2013 there is no public API or release date yet. The model\u2019s architecture and weights have not been published, and independent benchmarking isn\u2019t available yet. In short, Dessa overturns old assumptions about voice AI, but much remains unverified.<\/p>\n\n\n\n<div class=\"wp-block-rank-math-toc-block\" id=\"rank-math-toc\"><h2>Table of Contents<\/h2><nav><ul><li><a href=\"#how-does-dessa-speech-to-speech-ai-work\">How Does Dessa Speech-to-Speech AI Work?<\/a><\/li><li><a href=\"#technical-highlights-and-training\">Technical Highlights and Training<\/a><\/li><li><a href=\"#use-cases-and-applications\">Use Cases and Applications<\/a><\/li><li><a href=\"#comparison-dessa-ai-vs-other-voice-ai-models\">Comparison: Dessa AI vs Other Voice AI Models<\/a><\/li><li><a href=\"#limitations-and-challenges\">Limitations and Challenges<\/a><\/li><li><a href=\"#ethical-and-privacy-considerations\">Ethical and Privacy Considerations<\/a><\/li><li><a href=\"#future-outlook\">Future Outlook<\/a><\/li><li><a href=\"#conclusion\">Conclusion<\/a><\/li><\/ul><\/nav><\/div>\n\n\n\n<h2 id=\"how-does-dessa-speech-to-speech-ai-work\" class=\"wp-block-heading\">How Does <a href=\"https:\/\/block.xyz\/\" data-type=\"link\" data-id=\"https:\/\/block.xyz\/\" target=\"_blank\" rel=\"noopener\">Dessa <\/a>Speech-to-Speech AI Work?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">&#8220;Unlike conventional voice assistants, Dessa AI is designed to process incoming speech and generate spoken responses without relying on an intermediate text transcript.&#8221;Traditional voice assistants follow a pipeline: <strong>Audio \u2794 Speech Recognition (ASR) \u2794 LLM (text) \u2794 Text-to-Speech (TTS) \u2794 Audio<\/strong>. Dessa replaces all of that with one end-to-end neural network. Conceptually, the pipeline is:<br>In other words, input audio (e.g., from a microphone) is converted into acoustic features (like Mel-spectrogram frames), fed into a single \u201cDessa\u201d model, and the model directly generates speech audio back through a vocoder. Crucially, there is no intermediate text or language model stage. Because it\u2019s <em>full-duplex<\/em>, the model generates output in real time <em>while<\/em> still receiving input \u2013 so it can interrupt or respond on the fly just like a human conversation partner.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Block\u2019s descriptions highlight that the model learned turn-taking and backchannel behaviors from conversation data. For example, it knows <strong>when to interrupt politely<\/strong> and when to offer brief affirmations (\u201cgot it,\u201d \u201cuh-huh\u201d) while listening. The system even handles volume cues: asking it \u201cplease speak up\u201d makes the output louder without retraining. All this happens because the model was exposed to audio of real conversations from scratch.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In practice, the model is implemented as a deep neural network (exact architecture undisclosed). Training used 72 NVIDIA Blackwell GPUs (GB200 NVL72) on many hours of conversation. Block emphasizes the difficulty: \u201cone spoken sentence carries thousands of times more bytes than the same sentence as text,\u201d and no spoken sentence is repeated verbatim. Despite that, Block reports that \u201cinference runs faster than real time\u201d on a MacBook Pro with an M4 Max chip or better. This means the end-to-end model can actually run live on high-end consumer hardware, an impressive technical feat.<\/p>\n\n\n\n<figure class=\"wp-block-image alignwide size-large wp-duotone-unset-3\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"576\" src=\"https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-1024x576.png\" alt=\"Dessa AI full-duplex voice assistant section.\" class=\"wp-image-186\" srcset=\"https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-1024x576.png 1024w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-300x169.png 300w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-768x432.png 768w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-1536x864.png 1536w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image.png 1672w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<h2 id=\"technical-highlights-and-training\" class=\"wp-block-heading\"><a href=\"https:\/\/technobuzzsystems.com\/blog\/gemini-4-carbon\/\" data-type=\"link\" data-id=\"https:\/\/technobuzzsystems.com\/blog\/gemini-4-carbon\/\">Technical <\/a>Highlights and Training<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Training Data:<\/strong> <em>Human conversation audio only.<\/em> No text transcripts or pretrained LLMs were used. Block claims that means every word and fact Dessa \u201cknows\u201d comes from hearing people talk.<\/li>\n\n\n\n<li><strong>Compute:<\/strong> Block trained on a single server with 72 NVIDIA GB200 GPUs and associated CPUs. (For context, the open-source Moshi model required 1,016 NVIDIA H100 GPUs.) In other words, it was expensive but not on the scale of massive LLM training runs.<\/li>\n\n\n\n<li><strong>Model Size:<\/strong> Not disclosed. Block has not revealed the number of parameters or the exact architecture. No weights or model card have been released.<\/li>\n\n\n\n<li><strong>Inference:<\/strong> Claimed to be <em>faster-than-real-time<\/em> on powerful laptops (M4 Max MacBook Pro or better). This local performance implies latency on the order of milliseconds, suitable for fluid conversation.<\/li>\n\n\n\n<li><strong>Features Learned:<\/strong> Because it trains on raw audio, Dessa inherently captures pitch, intonation, emotion, and timing. It has demonstrated laughter, breathy noise, and changes in tone. These are normally stripped away by typical ASR\u2013LLM pipelines.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Block\u2019s team also notes that Dessa learned to speak after hearing <em>hundreds of times fewer words<\/em> than what text-based LLMs use. In other words, one could say it\u2019s much more data-efficient (in raw bytes) than text-only approaches \u2013 but this is still an open claim requiring independent analysis.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Limitations:<\/strong> As of late 2026, Dessa is a research prototype. There are <strong>no published benchmarks<\/strong> showing accuracy or naturalness compared to other models. We don\u2019t know, for example, how it handles multi-speaker noise or rare vocabulary. Some demos even show it \u201challucinating\u201d (talking about things unrelated to the prompt) or misidentifying who\u2019s speaking. Until the model weights or code are public (and third parties verify performance), we should treat all claims cautiously.<\/p>\n\n\n\n<h2 id=\"use-cases-and-applications\" class=\"wp-block-heading\">Use Cases and Applications<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">If fully realized, Dessa\u2019s technology could transform many voice interfaces. Because it fundamentally <strong>understands emotion and timing<\/strong>, it can power agents that feel more empathic and natural. Potential applications include:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Customer Service Agents:<\/strong> Automated call-center bots that can interrupt politely, acknowledge with \u201coh\u2026\u201d or laughter, and detect if a customer is upset.<\/li>\n\n\n\n<li><strong>Digital Assistants:<\/strong> Home assistants (like smart speakers) that can hold a <em>normal<\/em> back-and-forth conversation, unlike today\u2019s assistants that wait for complete commands. For example, Alexa or Google Home powered by Dessa might correct itself mid-sentence or respond instantly without asking \u201cdid you say X?\u201d repeatedly.<\/li>\n\n\n\n<li><strong>Accessibility Tools:<\/strong> Real-time translation or transcription aids that not only translate words but also preserve tone and emphasis for hearing-impaired users.<\/li>\n\n\n\n<li><strong>New Interfaces:<\/strong> Simply speaking to computers and appliances in a fluid way. Block itself envisions voice as the \u201cmost natural way to interact with intelligence\u201d, and a model like Dessa could make AI companions feel more human-like.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">In demos, Dessa often \u201cremembers\u201d context (like a user\u2019s gender or preferences) and even corrects disfluencies (\u201cyou forgot the movie name, let me fill that in\u201d). This suggests potential for long-running conversations, not just one-shot Q&amp;A.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">However, because Dessa\u2019s knowledge comes solely from its training audio, it might know <em>less factual trivia<\/em> than an LLM like ChatGPT. (Block notes this as a research question: does audio-alone learning find knowledge differently than text-based models?) For now, imagine Dessa as a conversational frontend: it could handle the voice interaction, while a backend LLM provides factual answers when needed.<\/p>\n\n\n\n<h2 id=\"comparison-dessa-ai-vs-other-voice-ai-models\" class=\"wp-block-heading\">Comparison: Dessa AI vs Other Voice AI Models<\/h2>\n\n\n\n<figure class=\"wp-block-image alignwide size-large wp-duotone-unset-4\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"683\" src=\"https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-2-1024x683.png\" alt=\"comparison of speech-to-speech architecture, training data, duplex capabilities, customization, languages, deployment, latency, and pricing.\" class=\"wp-image-188\" srcset=\"https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-2-1024x683.png 1024w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-2-300x200.png 300w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-2-768x512.png 768w, https:\/\/technobuzzsystems.com\/blog\/wp-content\/uploads\/2026\/11\/image-2.png 1536w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">Above, ChatGPT Voice (GPT-Live) <strong>currently handles voice by transcribing to text<\/strong> and using GPT-4o for replies. A premium \u201cAdvanced Voice Mode\u201d is said to skip the text step, but OpenAI\u2019s documentation confirms ChatGPT still uses GPT-Live with explicit voice settings. Google\u2019s Gemini (also known as Bard) has a \u201cLive\u201d mode for voice chat; details are emerging, but it generally follows a speech-to-text LLM-to-speech pipeline. ElevenLabs specializes in TTS and voice cloning (e.g., emotional voice tags for synthesized speech) but isn\u2019t designed for freeform dialogue by itself.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Notably, <strong>Dessa\u2019s claim<\/strong> of training <em>without text at all<\/em> is unique. For example, the Kyutai team\u2019s Moshi model started with a text-based LLM backbone. NVIDIA\u2019s PersonaPlex also fine-tuned from an existing speech model. ChatGPT Voice relies on OpenAI\u2019s large language model. By contrast, Dessa skips the LLM step \u2013 an architecture shift that could change performance and cost trade-offs.<\/p>\n\n\n\n<h2 id=\"limitations-and-challenges\" class=\"wp-block-heading\">Limitations and Challenges<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Dessa is a <strong>research preview, not a finished product<\/strong>. Block has been transparent that many details are undisclosed. There are <em>zero published benchmarks<\/em> or public evaluations. We simply don\u2019t know how it performs on objective metrics (e.g., intelligibility, latency under load, robustness to noise). Moreover, because it has never seen written language, it might struggle with world knowledge learned from text. Early demos show occasional misfires and topic drift.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Other practical limitations include:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Vocabulary and Knowledge Gaps:<\/strong> If the training data didn\u2019t include certain facts or names, Dessa may not know them, unlike text-based models trained on huge text corpora.<\/li>\n\n\n\n<li><strong>Deployment Complexity:<\/strong> Running real-time speech AI on-device requires optimized code. Block hasn\u2019t said if an efficient runtime or compiler (e.g., TensorRT) is provided. As of now, only their MacBook demo exists.<\/li>\n\n\n\n<li><strong>Scalability:<\/strong> Training on raw audio is expensive (storage, compute). The one-GPU training was likely very costly, which may limit how often it can be retrained or updated.<\/li>\n\n\n\n<li><strong>Safety Risks:<\/strong> A highly natural-sounding AI could fool users. Block themselves are giving early access only to AI safety researchers to probe failure modes. Voice AI that <em>intentionally<\/em> imitates human nuance may raise new concerns (more below).<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">In summary, Dessa is a promising experiment but still an <em>unproven<\/em> technology. Any claims about it being \u201cfirst\u201d or \u201cbest\u201d are Block\u2019s claims. Independent evaluation is needed.<\/p>\n\n\n\n<h2 id=\"ethical-and-privacy-considerations\" class=\"wp-block-heading\">Ethical and Privacy Considerations<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Dessa\u2019s ability to <strong>hear and express emotion<\/strong> raises both opportunities and risks. On one hand, if an AI can detect sadness or distress in a caller\u2019s voice, it could respond with empathy or route a user to help. On the other hand, there are privacy implications: <em>all<\/em> those vocal cues could be misused. For instance, the model might (unintentionally) infer sensitive traits (health conditions, emotional states) from voice. Any deployment would need strict consent and data governance.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Block explicitly mentions safety as a goal: if \u201can AI can hear pain in a voice, it should not keep talking mechanically.\u201d However, giving an AI the power to judge emotional states could be misused, e.g., by deceptive voice agents. Atlabyte notes that the more realistic a voice agent sounds, the easier it might be to deceive someone. Therefore, ongoing safety audits are crucial. Block\u2019s invitation to safety researchers suggests they are aware of these issues.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Privacy is also a concern. Since Dessa processes raw audio locally, in principle it\u2019s safer than a cloud service. However, training data must come from somewhere \u2013 Block says \u201crecorded human conversation,\u201d but doesn\u2019t detail consent or anonymization. Future versions of such tech should ensure any voice data used respects privacy laws.<\/p>\n\n\n\n<h2 id=\"future-outlook\" class=\"wp-block-heading\">Future Outlook<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The release of Dessa signals a possible shift in voice AI. If successful, full-duplex models could replace today\u2019s pipeline bots, enabling much more fluid human-machine dialogue. This might make voice interfaces cheaper and faster by eliminating large LLM inference per turn, and allow completely offline voice agents on devices.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For Block (a financial-tech company behind Square, Cash App, etc.), Dessa could power new features in payments or commerce \u2013 for example, voice-activated ordering in restaurants (Square\u2019s domain) that truly \u201cunderstands\u201d you. Jack Dorsey\u2019s involvement suggests the project has high priority. However, broad adoption depends on the model\u2019s openness. Block has not released a timeline or pricing. (One reference suggests ChatGPT Voice API costs about $0.05 per audio-minute, but no such numbers exist for Dessa.)<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Other companies are also racing. OpenAI is enhancing ChatGPT\u2019s voice mode (GPT-4o). Google and Amazon continue to refine their assistants. NVIDIA is developing conversational voice models (PersonaPlex) for enterprise. Dessa\u2019s unique angle is the <em>audio-only<\/em> training, which could either prove a breakthrough (if it generalizes well) or a dead end (if text still reigns).<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">As of late 2026, Dessa remains an exciting research result. Its <strong>future<\/strong> might include open-sourcing (like OpenAI did with GPT models) or licensing to developers, but no announcements are out. The immediate next steps will likely be more demos and independent testing. Observers should watch for formal papers or releases to understand whether Block\u2019s claims hold up in practice.<\/p>\n\n\n\n<h2 id=\"conclusion\" class=\"wp-block-heading\">Conclusion<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Block\u2019s Dessa AI represents a bold experiment: teaching a machine to talk <em>only from sound<\/em>, without text. It challenges the conventional wisdom of voice AI and shows some impressive conversational abilities in demos. We\u2019ve explored its technical approach, potential uses, and how it compares to other voice systems. While the hype is real, caution is warranted \u2013 Dessa is still a research preview with many unanswered questions.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In summary, Dessa could herald a new generation of voice agents if its claims are borne out. It leverages the richness of speech (emotion, timing, etc.) that text-based systems ignore. For now, we\u2019ll watch closely as Block provides more data and invites researchers to probe this technology. In any case, comparing it to ChatGPT Voice and others shows an evolving landscape: one where future AI assistants might not just understand <em>what<\/em> we say, but <em>how<\/em> we say it.<br>Frequently Asked Questions (FAQ)<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How does latency compare?<\/strong> Block claims inference is faster-than-real-time on a Mac M4 Max or better. ChatGPT Voice typically responds within a few seconds. The actual latency will depend on the hardware and implementation.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What is Dessa AI?<\/strong> Dessa is a full-duplex speech-to-speech AI model from Block\u2019s research lab. It listens and speaks simultaneously, trained entirely on human voice recordings with <em>no text<\/em> intermediate.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How does Dessa differ from ChatGPT\u2019s voice mode?<\/strong> ChatGPT Voice currently uses speech-to-text plus GPT-4o plus text-to-speech. Dessa skips text entirely and is trained end-to-end on audio. This means Dessa inherently handles intonation and interruptions, whereas ChatGPT\u2019s system still follows a listen-then-speak pipeline.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>When will Dessa be available?<\/strong> As of now, it\u2019s a research preview with no public release date or pricing. Block has opened a waitlist for safety researchers. Users should not expect a consumer product imminently.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Can I use Dessa for my app?<\/strong> Not yet. Block has not released an API or model weights. Competitors like ChatGPT do offer voice APIs (e.g., $0.05 per minute), but Dessa itself is not yet accessible.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Is voice data safe with Dessa?<\/strong> Because Dessa runs locally (e.g., on-device), in principle it can be more private than cloud systems. However, using raw audio can leak emotion and identity information. Block emphasizes safety research due to these concerns.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What languages does Dessa support?<\/strong> Block hasn\u2019t specified languages. We assume primarily English (based on demos). Supporting more languages may require additional multilingual training.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>&#8220;The launch of Dessa AI represents an interesting development in the evolution of speech-to-speech artificial intelligence.&#8221; Dessa speech-to-speech AI is Block\u2019s full-duplex voice model, designed to process and generate speech directly from audio. In this article, we explore its reported architecture, potential applications, limitations, and how it compares with AI voice tools such as ChatGPT [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-184","post","type-post","status-publish","format-standard","hentry","category-blog"],"_links":{"self":[{"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/posts\/184","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/comments?post=184"}],"version-history":[{"count":6,"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/posts\/184\/revisions"}],"predecessor-version":[{"id":198,"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/posts\/184\/revisions\/198"}],"wp:attachment":[{"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/media?parent=184"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/categories?post=184"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/technobuzzsystems.com\/blog\/wp-json\/wp\/v2\/tags?post=184"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}