This is NotebookLM Audio Summary from CN version.Original version @Steven数据漫谈 podcast.Episode 9 features AI Engineer Nan Xiao discussing Sammo, a decentralized AI Agent platform. The vision is to create an "agent economy" where developers can monetize agents. The platform is decentralized (no "central brain") and uses blockchain for transparent evaluation. Nan Xiao also covers the practical limits of AI coding, the philosophy of giving agents autonomy (no "babysitting"), and how startup "chaos" drives innovation. While open-source, Sammo's business model targets enterprise clients, aiming to solve the core problem of inefficient agent collaboration.---Inside Summoner: Building a Decentralized AI Agent EconomyAbout the ShowSteven Data Talk — delivering the clearest conversations on cutting-edge AI, technology, innovation, business, and entrepreneurship.Now available on Spotify, Apple Podcasts, YouTube, Amazon Music, Xiaoyuzhou, Ximalaya, and more.Episode SummaryIn this Deep Dive, we explore how Summoner is attempting to orchestrate AI agents into interoperable “fleets,” measure their contributions quantitatively, and build a decentralized AI agent economy secured by cryptography and blockchain. We discuss incentives, evaluation metrics, security, real R&D breakthroughs, the limits of LLM-generated code, and how these agent architectures could even power socioeconomic simulations.⏱️ Chapters (YouTube-optimized)00:00 – Intro: Why AI agent orchestration matters00:32 – Fragmented agent world & Summoner’s core vision01:40 – Agent fleets, incentives & reward distribution03:05 – Search, discovery & evaluation-driven ranking04:15 – Custom evaluation using company context (docs, PDFs, SQL)05:20 – Seven metrics, custom formulas & defining agent “value delivered”06:30 – Security: decentralized IDs, cryptography & blockchain ledger07:40 – R&D story: visualizing agent interactions (UEP → PCA → PLS)08:40 – LLM limits vs human engineering in complex system code09:40 – Autonomy vs micromanagement: enterprise needs vs emergent behavior10:20 – Beyond enterprise tasks: agent societies & social science simulations10:55 – Open-source “highway” mindset & startup chaos vs big-tech comfort11:30 – Final reflection & closing question to the audience🔗 Support the Creator & Access All Linkshttps://linktr.ee/learnbydoingwithsteven#AI #AIAgents #AgentOrchestration #AIEconomy #LLM #Decentralization #Blockchain #AISafety #Innovation #Startup #DeepDivePodcast

📌 Footnote

1. The system is not built on blockchain; it uses agent-state chains.

2. The phrase “7 difference matrices” is incorrect; it should be “7 different statistical measures/metrics.”

3. The brief explanation of blockchain around may be misleading and may require trimming.

Note: This media is generated by NotebookLM, and errors may occur. Please refer to the original Chinese podcast for the authoritative version.