{"generated_at": "2026-05-29T22:31:54.147386Z", "slug": "nvidia_ai_podcast", "source_id": "src_nvidia_ai_podcast", "name": "NVIDIA AI Podcast", "episode_count": 3, "avg_signal": 78.7, "median_signal": 83.8, "top_signal": 84.0, "latest_episode_at": "2026-05-27T15:45:00Z", "earliest_episode_at": "2026-05-13T13:00:00Z", "category_mode": "ai", "cover_image_url": "https://megaphone.imgix.net/podcasts/d5e0ea3c-5930-11f1-8aef-6fa48be07041/image/8a5e9e091e81bcdfeb1383a68af9f238.png?ixlib=rails-4.3.1&max-w=3000&max-h=3000&fit=crop&auto=format,compress", "rank_score": 106.262, "episodes": [{"episode_id": "ep_nvidia_ai_podcast_602979d4da6a", "episode_title": "Everyone Can Build a Robot: Open Source Embodied AI With Seeed Studio | NVIDIA AI Podcast Ep. 300", "podcast_name": "NVIDIA AI Podcast", "podcast_slug": "nvidia_ai_podcast", "source_id": "src_nvidia_ai_podcast", "category": "ai", "publish_date": "2026-05-27T15:45:00Z", "overall_score": 68.4, "score_breakdown": {"clarity": 75.0, "originality": 85.0, "hype_penalty": 3.0, "actionability": 65.0, "technical_depth": 54.0, "information_density": 58.0}, "podcast_cover_url": "https://megaphone.imgix.net/podcasts/d5e0ea3c-5930-11f1-8aef-6fa48be07041/image/8a5e9e091e81bcdfeb1383a68af9f238.png?ixlib=rails-4.3.1&max-w=3000&max-h=3000&fit=crop&auto=format,compress", "source_link": "https://ai-podcast.nvidia.com", "audio_url": "https://cohst.app/pdcst/9V4R8T/traffic.megaphone.fm/NVC9592437490.mp3", "listen_url": "https://cohst.app/pdcst/9V4R8T/traffic.megaphone.fm/NVC9592437490.mp3", "verdict": "worth_your_time", "why_listen": "Learn how open source hardware and natural language-controlled robotics are lowering barriers to building and deploying physical AI systems.", "summary": "Seeed Studio is democratizing robotics through open source hardware, offering a $200 open-source robot arm (SOAR) that integrates with NVIDIA's Jetson and OpenClaw for accessible embodied AI development. The company enables rapid prototyping and manufacturing, supporting a community-driven model where users\u2014from students to startups\u2014can modify, extend, and deploy physical AI systems. Real-to-sim workflows and end-to-end learning reduce the need for traditional coding, allowing intuitive robot training via demonstration."}, {"episode_id": "ep_nvidia_ai_podcast_6304755d881b", "episode_title": "Inside AI Tokenomics: How to Profitably Turn Tokens Into Business Value | NVIDIA AI Podcast Ep. 299", "podcast_name": "NVIDIA AI Podcast", "podcast_slug": "nvidia_ai_podcast", "source_id": "src_nvidia_ai_podcast", "category": "ai", "publish_date": "2026-05-21T16:00:00Z", "overall_score": 84.0, "score_breakdown": {"clarity": 90.0, "originality": 85.0, "hype_penalty": 2.0, "actionability": 85.0, "technical_depth": 78.0, "information_density": 72.0}, "podcast_cover_url": "https://megaphone.imgix.net/podcasts/1158b364-53ec-11f1-82f7-975ad2e37d6f/image/dd94cdf81ffe709a11457a4d46e0dd55.png?ixlib=rails-4.3.1&max-w=3000&max-h=3000&fit=crop&auto=format,compress", "source_link": "https://ai-podcast.nvidia.com", "audio_url": "https://cohst.app/pdcst/9V4R8T/traffic.megaphone.fm/NVC1708428495.mp3", "listen_url": "https://cohst.app/pdcst/9V4R8T/traffic.megaphone.fm/NVC1708428495.mp3", "verdict": "must_listen", "why_listen": "You\u2019ll learn how to align AI infrastructure decisions with business value by using cost per token as a north star metric and leveraging extreme co-design to scale efficiently.", "summary": "Token value is determined not just by model size but by the intelligence and interactivity of the output, with cost per token being a critical metric shaped by extreme co-design across hardware, software, and ecosystem. Agentic workloads increase token demand due to multi-turn, user-independent LLM calls, requiring ultra-low latency infrastructure. Business monetization strategies include direct token sales, AI-infused product enhancements, and value-added services, with pricing driven by production cost and customer willingness to pay."}, {"episode_id": "ep_nvidia_ai_podcast_40ba4027f448", "episode_title": "Snap\u2019s Secret to Processing 10 Petabytes a Day: GPU-Accelerated Spark | NVIDIA AI Podcast Ep. 298", "podcast_name": "NVIDIA AI Podcast", "podcast_slug": "nvidia_ai_podcast", "source_id": "src_nvidia_ai_podcast", "category": "ai", "publish_date": "2026-05-13T13:00:00Z", "overall_score": 83.8, "score_breakdown": {"clarity": 85.0, "originality": 94.0, "hype_penalty": 3.0, "actionability": 75.0, "technical_depth": 85.0, "information_density": 75.0}, "podcast_cover_url": "https://megaphone.imgix.net/podcasts/02d8a5d0-4e59-11f1-9037-e3af07f8ddc0/image/eacd84f9fe5549e01f19f072ae12e461.png?ixlib=rails-4.3.1&max-w=3000&max-h=3000&fit=crop&auto=format,compress", "source_link": "https://ai-podcast.nvidia.com", "audio_url": "https://cohst.app/pdcst/9V4R8T/traffic.megaphone.fm/NVC2926487857.mp3", "listen_url": "https://cohst.app/pdcst/9V4R8T/traffic.megaphone.fm/NVC2926487857.mp3", "verdict": "must_listen", "why_listen": "Learn how Snap scaled GPU-accelerated data processing to handle 10+ petabytes daily by repurposing existing online serving infrastructure and achieving dramatic cost and performance gains.", "summary": "Snap processes over 10 petabytes of data daily for experimentation, leveraging GPU-accelerated Apache Spark via Spark Rapids on NVIDIA GPUs to achieve 3.6x performance gains and reduce job costs by 76%. By creatively borrowing idle online serving GPUs and migrating workloads to Kubernetes, Snap flattened its data scale curve while maintaining reliability and SLAs."}], "category_breakdown": [{"category": "ai", "count": 3}]}