The Stack Overflow Podcast

Mark Cavage, President and COO of Docker and seasoned cloud infrastructure leader, talks hardened containers, sandboxes, and Docker's approach to securing AI agents. He explains hardened base images, MicroVM-based sandboxes, saving mutated environments, and observability and controls for agent workflows. Conversation covers migration tools, scaling implications, and Docker’s roadmap for agent-focused features.

Ryan sits down with Member of the Technical Staff at Anthropic and Model Context Protocol co-creator David Soria Parra to talk the evolution of MCP from local-only to remote connectivity, how security and privacy fit into their work with OAuth2 for authentication and authorization, and how they’re keeping MCP completely open-source and widely available by moving it to the Linux Foundation. Episode notes:The Model Context Protocol (MCP) is an open-source standard for connecting AI applications to external systems created by Anthropic. You can keep up with—or join—the work the MCP community is doing at their Discord server. Connect with David on Twitter. Today’s shoutout goes to Populist badge winner competent_tech for their answer to How do I review a PR assigned to me in VS 2022.TRANSCRIPTSee Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

Marcus Fontoura, Microsoft technical fellow and author of Human Agency in the Digital World, offers a clear take on technology, AI, and societal impact. He discusses algorithmic non-determinism and how social platforms distort information. He weighs efficiency against human dignity and argues for practical, human-centered uses of today's AI.

Thibault Sottiaux, engineering lead on Codex at OpenAI, builds agentic coding tools and secure SDLC tooling. He explains how agentic systems differ from chat assistants. He discusses dogfooding Codex inside OpenAI, sandboxing and safety, automated code reviews, onboarding agents to large codebases, multi-agent refactors, and ambitions for proactive, memory-enabled SDLC automation.

Philippe Saade, AI project lead at Wikimedia Deutschland who led the Wikidata Embedding Project, talks about vectorizing millions of Wikidata items for semantic search. They cover reducing scraping pressure, how items were transformed into searchable text, and combining vector search with SPARQL for discovery and precise queries.

Will Allen, Cloudflare VP who builds bot management and pay-per-crawl tooling. Josh Zhang, Stack Overflow SRE who runs bot traffic and DDoS defenses. Janice Manningham, Stack Overflow product leader on data licensing and access controls. They discuss why pay-per-crawl was needed, how crawlers and scraping evolved, technical 402 handling and tooling, and programmatic licensing vs enterprise deals.

Shireesh Thota, Corporate VP of Azure Databases at Microsoft, oversees SQL Server, Cosmos DB, Postgres and Fabric. He walks through cloud database evolution and architectures. Topics include Cosmos DB internals and indexing, Postgres as an extensible platform, disaggregated designs like HorizonDB, cost governance and multi-cloud data strategies, and what AI and vector/RAG features mean for future databases.

Scott Stephenson, CEO and co-founder of Deepgram and former particle physicist turned voice-AI leader. He discusses tackling speech recognition for noisy, real-world audio. They cover data vs. input features, synthetic audio generation, scalable low-latency streaming on cloud, responsible limits on voice cloning, and modular AI architectures for connected voice agents.

Tom Griffiths, Princeton professor bridging psychology and computer science, and author of The Laws of Thought. He traces logic from Aristotle and Boole to modern neural nets. Short takes cover why transformers learn language, how human inductive biases differ from LLMs, and what constraints might give machines conscious-like phenomenology.

Zuzanna Stamirowska, CEO of Pathway (building post-transformer, memory-first AI); Victor Szczerba, CCO at Pathway (enterprise product and observability lead); Rowan McNamee, Co-founder and COO of Mary Technology (legal fact-management for litigators). They discuss memory-inspired models and long attention spans, continual learning and efficient architecture, and using AI to extract, organize, and verify legal facts.