Writing
Technical deep-dives and thinking about where AI is headed.
The Capability vs Impact Gap In AI
Model capability far exceeds business impact. We look at why the gap persists and what might close it.
The Context Gap: Why AI Tools Underperform in Actual Work
LLMs often fail in real workflows less from lack of capability and more from missing state, norms, and permissions. Closing that gap is mostly product and integration work.
The Capability-Adoption Gap: Why AI Isn't Changing Everything Yet
AI models are remarkably capable. Business outcomes haven't caught up. We look at the data on where the gap exists, why it persists, and what might close it.
First Impressions of GPT-5.2: Early Tests in Our MCP-Based Eval Harness
Early observations from running GPT-5.2 through our MCP research setup. Too early for conclusions, but here's what we're seeing.
Setting Up MCP Servers: A Practical Guide for AI Development
A practical walkthrough of configuring MCP servers for Claude Code—from basic setup to debugging common issues.
Zero-shot player tracking with SAM3
How we built a vision system that tracks players across frames with persistent IDs, even through occlusions and camera cuts.