Software Engineering Daily

David Yanacek, Senior Principal Engineer at AWS who builds developer tools and agentic AI like Kiro. He talks about Kiro’s spec-driven AI IDE workflow that captures requirements, produces designs, and breaks work into testable tasks. Topics include spec versioning, property-based testing, steering files and powers, multi-agent coordination, and measuring agent effectiveness.

Krishna Sai, CTO at SolarWinds and veteran of enterprise IT and observability, discusses designing AI-driven, agentic systems for resilient operations. He talks about moving from monitoring to ambient agents, architecting LLM gateways and safe abstractions, compressing and summarizing operational data, and balancing autonomy with human-in-the-loop controls.

Jackson Sipp, a PhD researcher at the University of Colorado Boulder studying national-scale censorship systems. He explains how China’s Great Firewall worked during the 2021–2023 blocking window. Topics include the pop-count entropy detection technique, how padding and TLS-like headers were used to evade blocks, active probing and fingerprinting, and implications for circumvention tool design and future censorship trends.

Gideon Mendels, Co-founder and CEO of Comet who built ML systems at Google, talks about building and evaluating LLM-powered agents. He covers why non-determinism breaks traditional testing. He explains evals as test suites, bootstrapping regressions from production, and treating prompts, tools, and configs as optimization/search problems.

Steve Yegge, veteran software engineer and writer who builds experimental agentic tooling like Beads and Gastown. He discusses the shift from chat-style coding to orchestrating fleets of AI agents. Topics include task graphs and Git-backed shared memory, multi-agent coordination patterns, trust and scaling, and how workflows and roles change as agents take on more work.

Łukasz Langa, CPython Developer in Residence and long-time release manager, walks through Python 3.14’s big changes. He covers free-threaded no-GIL support and when it helps versus sub-interpreters. He explains template string literals, deferred type annotation evaluation, and new debugging and profiling tools. The conversation highlights performance, scalability, and usability improvements.

Adam Miskiewicz, a principal software engineer at Airbnb who led work on Viaduct, discusses Airbnb’s unified data graph and GraphQL platform. He recounts Viaduct’s origins, the architectural choices that enabled scaling to massive query volumes, and the reasons Airbnb open-sourced Modern Viaduct. They also touch on how typed, centralized data layers may shape backend platforms and AI-driven development.

Rapid in-flight internet and how it changes travel productivity. Tesla moving from Autopilot toward full self-driving. Apple reportedly using Google’s Gemini for Siri. Meta’s acquisition of Manus and Waymo’s scaling challenges. The intense arms race for AI talent and shifting stock compensation strategies.

Ed Bayes, product designer on Codex at OpenAI, shapes UX for coding agents. Thibault Sottiaux, Codex engineering lead, builds agent architectures and harnesses. They discuss agentic coding systems, sandboxing and safety trade-offs, co-evolution of models and harnesses, integrations across IDEs and toolchains, multi-agent futures, and latency and performance trade-offs.

Fred Roma, SVP of Product and Engineering at MongoDB, a veteran in cloud and data management. He talks about the complex AI stack: LLMs, embeddings, vector search, caching, and observability. He covers schema evolution in the LLM era, Voyage AI’s multimodal embeddings and rerankers, and how data platforms must adapt for production-grade AI systems.