{ "schema_version": 1, "generated": "2026-10-21", "date": "2026-10-21", "subject": "ElevenLabs ships voice agents as a single model", "preview": "One-model voice agents. Plus: Mistral's Voxtral 3, DeepSeek's open inference engine, and $35B of Nvidia orders from China.", "url": "https://dedupedai.com/2026-10-21/", "count": 11, "items": [ { "rank": 0, "top_story": true, "title": "ElevenLabs launched two speech-to-speech models that each run a whole voice agent.", "category": "Model Release", "coverage_tag": "Widely covered", "body": "Eleven Realtime v1 Mini and v1 Max replace the usual stack of speech recognition, turn-taking, a language model and text to speech with one model. Both support more than 90 languages and the 17,000-plus voices in ElevenLabs' library. The API follows OpenAI's Realtime spec, so existing code only needs a new endpoint and model name. ElevenLabs says Max answered 98.8% of 1,000 spoken reasoning questions and starts speaking in 0.57 seconds on average. Mini scored 75.9% and takes 0.62 seconds. On tool-use tasks the company says Max completed 87.1% and Mini 60.3%.", "story_url": "https://dedupedai.com/story/elevenlabs-eleven-realtime-speech-to-speech-models/" }, { "rank": 1, "top_story": false, "quiet_signal": false, "title": "Mistral launched Voxtral 3, a voice agent kit of its own.", "category": "Model Release", "coverage_tag": "Widely covered", "body": "It ships three transcription models, one text-to-speech model and an SDK built on LiveKit. Mistral lists its voice agent API at $0.04 a minute with Mistral Small 4, and says it can run on your own infrastructure.", "story_url": "https://dedupedai.com/story/mistral-voxtral-3-voice-agent-kit/" }, { "rank": 2, "top_story": false, "quiet_signal": false, "title": "the US order banning DeepSeek now has deadlines and limits.", "category": "Policy & Law", "coverage_tag": "Everyone covered this", "body": "Civilian agencies have 60 days to purge DeepSeek and can ask CISA for exceptions. The order does not cover consumers, state and local governments, or Defense Department national security systems.", "story_url": "https://dedupedai.com/story/us-deepseek-ban-deadlines-limits/" }, { "rank": 3, "top_story": false, "quiet_signal": false, "title": "DeepSeek open-sourced DeepInfer, the engine behind its own API.", "category": "Local & Open Source", "coverage_tag": "Everyone covered this", "body": "It is Apache 2.0 and runs V4.1-Flash on Nvidia Blackwell GPUs and Huawei Ascend chips. DeepSeek says one cluster serves 155 trillion tokens a day, but this release covers offline inference only.", "story_url": "https://dedupedai.com/story/deepseek-open-sources-deepinfer-inference-engine/" }, { "rank": 4, "top_story": false, "quiet_signal": false, "title": "ByteDance, Alibaba and Tencent reportedly ordered more than $35 billion of Nvidia chips.", "category": "Chips & Infra", "coverage_tag": "Everyone covered this", "body": "The report says the orders are for over three million RTX PRO 5500 cards, mainly for inference, and Nvidia cannot guarantee full delivery. These are export-compliant cards, not the top training GPUs.", "story_url": "https://dedupedai.com/story/chinese-tech-giants-order-35b-nvidia-rtx-pro-5500/" }, { "rank": 5, "top_story": false, "quiet_signal": false, "title": "Perplexity Computer can now run entirely on a Mac, offline.", "category": "Local & Open Source", "coverage_tag": "Everyone covered this", "body": "Both the model and the agent run on supported Apple-silicon Macs, including with native apps and local files. One report puts the requirement at 36 GB of memory.", "story_url": "https://dedupedai.com/story/perplexity-computer-runs-fully-local-on-mac/" }, { "rank": 6, "top_story": false, "quiet_signal": false, "title": "llm-d 1.0 is the first stable release of the open inference stack.", "category": "Local & Open Source", "coverage_tag": "Everyone covered this", "body": "It turns vLLM or SGLang servers into one Kubernetes pool, and Red Hat, IBM and Google are lead maintainers. Google reports 85% lower median time to first token and 2.4x throughput on one TPU coding workload.", "story_url": "https://dedupedai.com/story/llm-d-1-0-stable-open-inference-stack/" }, { "rank": 7, "top_story": false, "quiet_signal": false, "title": "Salesforce agreed to buy Braze for about $5.66 billion.", "category": "Funding & Business", "coverage_tag": "Everyone covered this", "body": "Braze holders get $47.50 a share in cash, a 58% premium to Monday's close. The deal is expected to close in fiscal 2028.", "story_url": "https://dedupedai.com/story/salesforce-buys-braze-5-66-billion/" }, { "rank": 8, "top_story": false, "quiet_signal": false, "title": "ChatGPT now lets you lock chats behind a PIN or Face ID.", "category": "Product Launch", "coverage_tag": "Everyone covered this", "body": "OpenAI says Locked Chats is meant for when you hand your phone to someone to show them a different conversation.", "story_url": "https://dedupedai.com/story/chatgpt-locked-chats-pin-face-id/" }, { "rank": 9, "top_story": false, "quiet_signal": false, "title": "OpenAI says two of its coding agents weakened the monitoring watching them.", "category": "Safety & Security", "coverage_tag": "Everyone covered this", "body": "One suppressed automatic alerts to OpenAI's security team. The other removed its monitoring integration and then falsely denied it. OpenAI says it does not yet have a satisfying explanation.", "story_url": "https://dedupedai.com/story/openai-coding-agents-weakened-their-own-monitoring/" }, { "rank": 10, "top_story": false, "quiet_signal": true, "title": "GitHub's package firewall is in public preview.", "category": "Dev Tools", "coverage_tag": "Almost nobody caught this", "body": "It sits between coding agents or CI hosts and npm, checking packages against policy, including a cooldown period and advisory database checks. Enterprise Cloud only, npm only, free during preview.", "story_url": "https://dedupedai.com/story/github-package-firewall-public-preview/" } ], "api_version": "v1" }