Presenting the AI Engineer World's Fair — with Sam Schillace, Deputy CTO of Microsoft

Episode Summary

Executive Summary: This episode promotes the upcoming AI Engineer World's Fair while recapping the origin and growth of the AI Engineer Summit. The hosts and Microsoft’s Sam Schillace discuss why vendor-neutral, in-person events matter, how AI engineering is shifting from research to production, and how code plus LLMs should be combined: think with the model, plan with code. They also cover Microsoft’s broad AI portfolio and Schillace’s hands-on philosophy for building in a fast-moving field.

Main Topics: Origin and growth of AI Engineer Summit: The hosts explain how the AI Engineer Summit emerged from early conversations in 2023 and became a major gathering for AI builders, with strong in-person and online attendance. Vendor-neutral conference philosophy: A major theme is why the event is intentionally vendor-neutral: to improve content quality, broaden the audience, and create a community-driven industry conference rather than a company showcase. AI Engineer World's Fair as a larger expo-style event: The new World's Fair is framed as a bigger, more ambitious in-person expo inspired by historical world's fairs, meant to showcase AI progress, foster networking, and slow down the online news cycle. Sam Schillace’s AI experience and Microsoft perspective: Schillace shares his early GPT-4 reactions, his work on Semantic Kernel, and how Microsoft is organizing AI efforts across research, product, and developer tooling. How to build with LLMs: code and semantics: Schillace argues that models are best used for thinking and semantic generation, while code should handle planning, orchestration, and reliable execution in longer-running systems. Multimodality, agents, and the limits of current systems: The conversation touches on multimodal models, agentic systems, memory, stochasticity, and why current autonomous systems can get stuck or drift without strong code scaffolding. Microsoft’s AI portfolio and leadership style: Schillace describes Microsoft as spanning research, infrastructure, and product surfaces, and emphasizes Satya Nadella’s unusually hands-on leadership and the company’s all-in AI posture.

Key Arguments: Vendor-neutral events attract a broader, more technical audience and produce better content than company-owned conferences. The AI Engineer Summit succeeded because it captured a real emerging industry and gave it a place to congregate. The World's Fair format is meant to create an in-person, expo-like environment that supports deeper conversations and community building. GPT-4 felt like a category shift: from syntax-heavy computing to semantics-heavy interaction. LLMs are better at thinking than planning; code should provide the exoskeleton that makes AI systems reliable. Long-running agent systems are vulnerable to stochastic noise, memory issues, and feedback loops, so orchestration must be explicit. Microsoft’s AI strategy spans research, developer tools, cloud infrastructure, and end-user products, making it both a platform and a consumer-facing AI company. Hands-on experimentation is essential for leaders; opinions about AI are not enough without building and breaking things directly.

Data Points: AI Engineer Summit in-person applications: over 5,000 - Number of people who applied to attend the first summit in person. AI Engineer Summit online attendance: over 9,000 - Total online attendees for the summit. Live stream audience: over 20,000 - People who watched the summit live stream. In-person attendance accepted: about 400 - Approximate number of attendees admitted in person. Event duration: two days, two and a half days with a welcome reception - Length and structure of the original summit. World's Fair tracks: 9 - Number of tracks announced for the AI Engineer World's Fair. Talks and workshops: over 100 - Total sessions planned across the nine tracks. Microsoft startup credits: up to $150,000 over four years - Founders Hub credits mentioned as support for startups. Startup entry cost: about $2,500 - Approximate low-cost starting point mentioned for Microsoft startup support. CFP deadline: April 8 - Deadline for speaking proposals at the World's Fair. Event dates: June 25, 26, and 27 - Dates for the AI Engineer World's Fair in San Francisco.

Pivotal Quotes: "Think with the model, plan with code." — Sam Schillace: His core framework for combining LLMs with deterministic software orchestration. "This is really like the beginning, it's the beginning of V2 of the computer industry completely." — Sam Schillace: His reaction to GPT-4 as a category-shifting technology. "If you want the truly internet community, if you want a more varied audience and, frankly, better content for, especially for a technical audience, you want a vendor-neutral event." — Ben Dunphy: Why the conference is intentionally not owned by a single company.

Implications: The episode signals that AI engineering is maturing into a real industry with its own flagship gatherings, tooling patterns, and leadership practices. For builders, the message is to combine LLM reasoning with robust code, and for companies, to engage hands-on rather than from the sidelines.