Research
What we're discovering through research
Research, analysis, and ongoing inquiries into AI middleware, agentic systems, developer tooling, and the evolving relationship between foundation models and the systems around them.
What we focus on
Our research emerges from building and operating middleware in production. These are the areas where that vantage point gives us observations that benchmarks and controlled environments often miss.
Agent Reliability
Understanding how agents behave in production environments: tool failures, coordination breakdowns, recovery strategies, and long-running task reliability.
AI Economics
Studying the cost, latency, routing, and infrastructure decisions that determine whether AI systems create value at scale.
Human-AI Collaboration
Investigating escalation patterns, human oversight, trust calibration, and the boundaries between automation and human judgment.
Middleware Systems
Exploring orchestration patterns, protocol design, tool integration, and the systems that connect models to real-world software.
Research Papers
Our research is grounded in the systems we build and operate. We document what works, what fails, and what those observations mean for builders.
How we researchSignal, not noise.
Focused newsletter for builders and knowledge workers tracking how AI is changing real work. We surface what matters in practice, not every headline. Curated for practitioners, not spectators.