Latest episode

The Outpost Ep 7- MCP 101, Biology of an LLM & Physical AI by Soham, Dushyant & Saranya

Welcome back to The Outpost! In this episode, Soham unpacks the Model Context Protocol (MCP) — the emerging standard that's transforming how LLMs interact with the world, making them not just smarter, but actionable. We explore how MCP could reshape AI deployment, agent ecosystems, and even the future of app interfaces. Then, Saranya does a deep dive into Physical AI — where intelligence meets the real world. From sandwich-making bots to smart fabrics and robotic adaptation, we cover key breakthroughs and the regulatory challenges ahead. Finally, Dushyant takes a peek inside the LLM "brain": How do these models really think, rhyme, translate, and sometimes hallucinate? What does it mean for explainability and trust? Timestamps below 👇 00:00 – Intro & Agenda 00:36 – MCP: Making LLMs Useful 03:14 – How MCP Works & Why It Matters 07:08 – Agents, APIs & Real-World Impact 10:06 – MCP Ecosystem & Ethical Concerns 12:46 – What Is Physical AI? 18:34 – Embodied vs Physical AI 19:47 – Adaptability & Real-World Challenges 22:42 – Key Breakthroughs (MIT, Fagel) 26:09 – Risks, Regulation & the Road Ahead 30:30 – Inside the LLM Brain 34:14 – Thinking, Reasoning & Rhyming 42:29 – Multilingual & Math Capabilities 47:51 – Hallucinations & Trust Issues 51:17 – Explainability & Closing Thoughts Links MCPs 1. Sequoia Capital: AI’s $600B Question: https://www.sequoiacap.com/article/ais-600b-question/ 2. Anthropic: Introducing the Model Context Protocol: https://www.anthropic.com/news/model-context-protocol 3. MCP explained: MCP Explained: The New Standard Connecting AI to Everything https://medium.com/@elisowski/mcp-explained-the-new-standard-connecting-ai-to-everything-79c5a1c98288 4. Andreesen Horowitz: A Deep Dive Into MCP and the Future of AI Tooling: https://a16z.com/a-deep-dive-into-mcp-and-the-future-of-ai-tooling/ 5. MCP.so: MCP Server https://mcp.so/ 6. Manus AI: https://manus.im/ Physical AI 1. What is Physical AI? https://www.techtarget.com/whatis/feature/Physical-AI-explained-Everything-you-need-to-know 2. Smell sensors: https://cen.acs.org/analytical-chemistry/chemical-sensing/new-artificial-nose-like-sensor/102/i16 3. Gemini's Physical AI models: https://www.wired.com/story/googles-gemini-robotics-ai-model-that-reaches-into-the-physical-world/ 4. Liquid Neural Networks:https://news.mit.edu/2023/drones-navigate-unseen-environments-liquid-neural-networks-0419 5. FaGeL: https://arxiv.org/abs/2412.20297 Biology of an LLM 1. On the biology of a large language models: https://transformer-circuits.pub/2025/attribution-graphs/biology.html 2. Circuit tracing: revealing computational graphs in language models: https://transformer-circuits.pub/2025/attribution-graphs/methods.html Visit Our Website: https://www.theoutpost.ai Support Us: Like, Comment, Share and Subscribe. Your engagement helps us grow!