Signals

Discussion on the challenges and limitations of using LLMs for code generation in production environments, exploring determinism concerns.

Discussion of advanced techniques for constraining LLM outputs using structured generation libraries like Guidance and LLGuidance.

How BlackRock's engineering team overcame challenges in building and deploying AI applications at scale for investment operations teams.

How to build reliable AI systems while navigating the rapid pace of LLM evolution and the tension between domain knowledge and scalable methods.

Key insights on building robust evaluation systems that enable rapid model updates and continuous improvement in AI applications.

Discussion on how complex algorithmic problems can be simplified using constraint solvers, and the trade-offs between custom solutions and dependency management.

A systematic approach to collaborating with AI systems on software projects, addressing code bloat, architectural drift, and context dilution through structured constraints and validation checkpoints.

Learn how to implement OAuth2 authentication with dynamic client registration and authorization server metadata for MCP servers using a reverse proxy gateway approach.

Kent Beck explores the distinction between 'vibe coding' and 'augmented coding' through building a production-ready B+ Tree library with AI assistance.

Discussion on how AI coding tools might revitalize XP practices, with insights from Kent Beck on test-driven development and feedback loops.