Paul Veugen

Paul Veugen

Founder of Subwave & Detail. A glimpse behind the scenes, building a new platform to share your best stories with video and audio.

Episodes

  1. 3d ago ·  Video

    Little brains in every product

    We carry more compute in our hands than sits in the entire cloud, yet almost all the money flows to data centers. For five years we've built on-device first at Detail and Subwave, shipping local rendering, captions, and enhancement while keeping video on the device. Through NP-Hard we backed Mirai, pushing LLM inference to 1,000 tokens per second, and at Detail we use Argmax's on-device speech models. The building blocks exist, but there's still a wide gap between what developers want to build and what's actually available to them. Every developer I spoke to at WWDC had the same story: a wish list of AI features they'd build if inference cost weren't a factor, and expensive cloud tokens they'd happily swap for a local model running on the device already in their user's pocket. The demand is real. The SDKs aren't there yet. That's why we're launching Desert Ant Labs, a European on-device AI lab focused on shipping dozens of small, opinionated audio and visual models that drop into any product with a few lines of code. No inference cost, nothing leaving the device, running on the 6 billion devices people already own. When cost drops to zero, you stop trading capability against budget on every feature decision. The first models already power Detail and Subwave, and a dozen more will be available to third-party developers before the end of the year. Published on Subwave https://subwave.app/@paul/post/little-brains-in-every-product

    2 min
  2. May 24 ·  Video

    Designing on-device AI models

    Two new on-device AI audio models are shipping across Detail and Subwave this week, and both were built in-house. Until recently, the benchmark for audio enhancement was Dolby: server-side processing, per-minute pricing, and a generic output that sounded identical regardless of the app it came from. Building the alternative became possible because of four things arriving at the same time: open-source base models, years of Detail recordings as training data, affordable hardware fast enough to train overnight, and tools like Claude Code handling the engineering scaffolding around the model itself. The Clear model cleans up recordings and the Uhm model detects filler words, both running entirely on-device, processing a 10-minute recording in 10 seconds without the audio ever leaving your phone. The audio enhancement model was shaped around a specific sound target: warm and present, closer to a podcast studio than a phone call. The clean reference audio it learns from is itself lightly enhanced and normalized toward that target, so the model learns the sound of Detail recordings, not clean speech in general. Iteration looked more like product work than research: train overnight, listen back in the morning across tens of real recordings, run blind A/B comparisons against Dolby and against previous versions, pick the winner, repeat. Because both models carry zero variable cost, they change what's possible as product defaults. In Detail, Auto Edit runs both models on every recording without being asked. In Subwave, audio enhancement applies to every post by default. The filler word detection model processes a 57-minute interview in under 30 seconds. The result is not a generic enhancement layer dropped into the app; it is a trained opinion about what a Detail recording should sound like. Published on Subwave https://subwave.app/@paul/post/designing-on-device-ai-models

    5 min

About

Founder of Subwave & Detail. A glimpse behind the scenes, building a new platform to share your best stories with video and audio.