{"slug": "voice-is-underutilized", "title": "Voice is Underutilized", "summary": "A practitioner reports that speech input and output remain underused with AI systems, citing a personal workflow that pairs NVIDIA's Parakeet STT models with the FluidVoice app and a Claude Opus cleanup pass to make local transcription nearly indistinguishable from 100% accurate. The practitioner also built a daily 10-20 minute text-to-speech briefing podcast, which required months of tuning to add schedule variety and avoid becoming stale, and uses transcribed design syncs and an onsite to preserve context for later agent sessions.", "body_md": "Speech is a useful medium for communicating with AI, both as an input/output channel and as back-and-forth. Over 2026 I’ve been increasing the amount of speech I/O and have collected a fair number of use cases.\n\nVoice is, of course, great for rambling at coding agents. Oftentimes the transactional “just talking with the agent” convo stuff is easier to ramble out than it is to type. I also occasionally have wrist sensitivity issues that make it difficult to type – now with STT and computer-use, that’s not a hassle. (Knock on wood about RSI 😬)\n\nVoice is also good for capturing context. Every phone has a voice recorder on it; most meeting software now has transcription support. There’s a lot of ambient context that just exists in conversations that, if you can (consentingly) capture it, has a ton of downstream value. For example: my team has been transcribing design syncs consistently. A few solo agent sessions interspersed between transcribed design syncs allow engineers to iterate on designs quickly without having to do the “sit down for several hours and write a doc” schtick.\n\nEarlier in the year, my team had an onsite where we got a few folks into the room and hammered out long-term plans. Those conversations were fully transcribed and have been instrumental in a bunch of the work I ended up doing later in the year. Having the context in textual form “froze” those ideation sessions, allowing me to rehydrate them months later, when I had cycles to think about the problems discussed again. That being said, there are obviously some things that you don’t want recorded, for one reason or another. In practice, it’s much more common for us to regret not having transcribed a meeting than the other way around.\n\nVoice input also allows you to write thoughts or capture things that you\nwouldn’t otherwise have time to do. For instance, I often now ramble into the\nVoice Memos app on my phone while commuting home about a problem I’m thinking\nabout, with a few proposed sketches of a solution – just pure\nstream-of-consciousness. Then when I get home, I transcribe the audio, set an\nagent up on it with `/goal` overnight, and come back to find (surprisingly\noften) the problem fully solved.\n\n**Models**: Local STT models are excellent. I’ve been using\n[FluidVoice](https://github.com/altic-dev/FluidVoice), with NVIDIA’s\n[Parakeet models](https://catalog.ngc.nvidia.com/orgs/nvidia/-/collections/parakeet-tdt-0.6b/-),\nand it’s remarkably accurate. Claude/ChatGPT are good at cleaning up\ntranscription errors. Parakeet + FluidVoice + a Claude Opus cleanup is nearly\nindistinguishable from 100% accurate for my use cases.\n\nSince early this year, I’ve been building and tweaking a personal daily briefing “podcast” which aggregates a bunch of information sources I care about, and uses a text-to-speech model to make a daily audio briefing.\n\nThis briefing contains things like:\n\nIt’s generally 10-20 minutes of audio daily. I tweaked it for a couple months to get it exactly how I like it, and now the quality is consistently high. The crucial, and most difficult, part of this was building enough variety into the “schedule” of what segments show up in the briefing every day that it doesn’t end up sounding stale, gimmicky, or formulaic.\n\nA simple “Claude, give me a daily briefing” becomes stale after a few days. You really do need entropy and novelty and continuously updating/streaming context for this sort of thing to be consistently interesting.\n\nBut now it *is* quite interesting to me! It’s made me significantly more aware\nof local news, and has allowed me to “read” more of the RSS feeds that I\nsubscribe to – especially articles which I likely wouldn’t have read due to the\ntitle not being interesting, but which the LLM thought were interesting based on\nits actual contents and knowledge of my interests.\n\nOutput is also handy for learning things with LLMs. Earlier this year I set up a system for generating “explainers” as podcasts, and learned quite a bit about aviation, various distributed systems topics, and other areas of interest through passive audio consumption.\n\n**Models**: I had been using GPT-4o mini TTS as my “programmatic output voice\nmodel of choice” for quite a while until its deprecation was announced\n([RIP](https://developers.openai.com/api/docs/deprecations#2026-10-01-text-to-speech-models)).\nI’ve now switched to\n[Gemini 3.8 TTS](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-text-to-speech/),\nand it’s *fantastic*. It features expressive voices, works over OpenRouter, and\nis cheap. Highly recommended.\n\nClaude, ChatGPT, and Gemini all have a “live” interactive voice mode – where you can conversationally talk with the model, and it can respond and take actions using tools, and so on. The primary time I use the interactive mode is during commutes. It’s helpful to use the voice mode to triage tasks, ask research questions, and iterate on ideas. It’s like taking time that would otherwise be dead time and translating it into something that can be productive.\n\nVoice is also convenient for walking around the house and capturing things that need to happen. As you remember things, you can have the AI make tasks or calendar entries and burn through a bunch of administrivia without interacting with your phone. It’s like having an actual, literal assistant.\n\nMy preferred model is Claude for voice interactive mode (even though it’s not a\n“native” voice-to-voice model like ChatGPT). A while back the Claude voice mode\nadded support for using MCPs, and that was a game changer.<sup>[1](#fn:1)</sup> I’ve built\na whole [personal set of software](https://benjamincongdon.me/blog/2026/11/02/Interpolative-Software/)\nthat has everything from tasks to notes to flashcards, all of which is\naccessible through MCP. The ability to use MCP via voice means that I can use\nthe live voice mode for triaging or quickly recording tasks.\n\nI also use it for reviewing flashcards. Since my MCP supports flashcards, I can have Claude ask me questions, I have to give the answer, and then the model grades my answer and records it. This is, in my opinion, strictly better than the normal Anki app, because with Anki, you can flip through cards quickly and semi-trick yourself into thinking that you actually know the answer. But if you’re forced to say it out loud: one, doing it verbally can help retention, in my opinion. And two, you are actually being graded. You’re not just in your mind thinking, “Oh yeah, I know how to do this.”\n\nWhen the models got access to web search through the voice mode, for months all I was doing was asking research questions about history or technology or anything that I was interested in, and having conversations about it. It’s a fun rabbit hole tool.\n\nBack in the GPT-4o days, you had to take anything it said with many grains of salt, and you still have to do that now. But if you use Opus, you can generally trust what it says in the voice mode, even though the voice modes tend to still lag behind the normal text modes in features, and you tend to be stuck on the older, worse models for a couple weeks before they upgrade them, even when everything else in the product is seemingly over on the new models.\n\n**TL;DR**: voice is underutilized! It’s a natural modality, and it’s\n“programmable” at this point. I’d suggest trying, if nothing else, the “on my\ncommute I record Voice Memo prompts or talk with the model about\ntasks/interests/etc.” I’ve already converted a few people to this and have\nreceived positive reactions. Make your own bespoke personal audience-of-one\npodcast, listen to articles you wouldn’t have time to read, record a brain dump\nand then have an LLM help you pick through it. There’s a ton of mundane utility\nto be had.\n\nI recall ChatGPT added this support recently too, but in my testing it was\nsomewhat inconsistent, and I had already standardized on using Claude for\nthis use case, which has been sticky. [↩︎](#fnref:1)", "url": "https://wpnews.pro/news/voice-is-underutilized", "canonical_source": "https://benjamincongdon.me/blog/2026/11/03/Voice-is-Underutilized/", "published_at": "2026-11-04 05:00:00+00:00", "updated_at": "2026-11-04 05:50:03.122352+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "ai-tools", "natural-language-processing"], "entities": ["FluidVoice", "NVIDIA", "Parakeet", "Claude", "ChatGPT", "Claude Opus"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/voice-is-underutilized", "markdown": "https://wpnews.pro/news/voice-is-underutilized.md", "text": "https://wpnews.pro/news/voice-is-underutilized.txt", "jsonld": "https://wpnews.pro/news/voice-is-underutilized.jsonld"}}