🚀 Livepeer Hits Record AI Fees

Mike Zupper's career spans the evolution of the internet era, bridging digital innovation with real-world applications across multiple sectors. **Career Highlights:** - Pioneered internet integration in politics, e-commerce, and healthcare - Transitioned into crypto and AI development - Known for his unconventional approach: constantly reinventing and starting fresh Zupper's path demonstrates how early internet builders adapted their skills to emerging technologies like blockchain and artificial intelligence. His willingness to "rip it up and start again" reflects the iterative nature of tech innovation. Read the full builder spotlight: [Mike Zupper's Career Story](https://livepeer.org/blog/builder-spotlight-mike-zupper)

**BlueClaw**, an OpenAI-compatible platform, has processed over **1 billion work units** on Livepeer's distributed GPU network. Built by Mike out of frustration with OpenAI's pricing model, the platform offers: - **No rate limits** for AI job processing - **Full OpenAI compatibility** - **Distributed GPU infrastructure** via Livepeer The milestone demonstrates growing adoption of decentralized alternatives to centralized AI platforms. BlueClaw leverages Livepeer's existing video infrastructure network to provide accessible, scalable AI compute resources without the restrictions of traditional providers.

Livepeer demonstrated new capabilities on their Water Cooler session, introducing **agent-compatible APIs** for their decentralized video network. **Key features:** - APIs designed for AI agents to discover and call autonomously - Transcription and computer vision services - Running on open, distributed compute infrastructure This builds on Livepeer's vision to create a **real-time video AI layer** powered by community infrastructure rather than closed cloud platforms. The open approach aims to provide the speed and scale needed for next-generation AI video applications.

Raid Guild is working on improving the delegator experience on Livepeer through a comprehensive UX analysis project. **Key Focus Areas:** - Analysis of delegator user experience - Comparison of self-custody vs custodial staking options - Streamlining staking flows to make delegation more effortless - Improving observability of the delegation process The community can track progress through multiple channels: - [Forum discussion](https://forum.livepeer.org/t/rfp-delegator-ux-analysis-self-custody-vs-custodial-staking/3254/6) - Discord updates - Weekly Water Cooler meetings (Thursdays at 2pm ET) This initiative aims to make the delegation process on Livepeer's video infrastructure more accessible and user-friendly for all participants.

**DaydreamLiveAI** released **Scope v0.2.5**, a 7-node workflow that generates real-time agent avatars with natural gestures, eye movement, and facial expressions. **Key features:** - Combines LTX Studio's LTX 2.3 model (rare open-weights model) - Joint audio-video generation from single prompt - Face-front avatar LoRAs trained on close-up footage - Natural lip sync, eye motion, and microexpressions - Runs on single GPU **Performance improvements:** - Skips CPU offload on high-VRAM GPUs - Faster generation speed vs. previous flat talking faces The workflow addresses a common limitation where image-to-agent avatars typically appear as static talking heads, instead delivering dynamic, responsive avatars that react to their own lighting. [View announcement](https://x.com/DaydreamLiveAI/status/2055269758891630799?s=20)