Giving AI the Gift of Memory and Context at Silicon Harbor Technologies

Join host Bill Billotte as he sits down with Kenny Alge and Jake Mundell, co-founders of Charleston-based Silicon Harbor Technologies, to dive deep into their company's game-changing pivot. What started as human-facing tools like DataVerse and Explorer evolved into a profound question: What if the real consumer of code intelligence isn't developers... but AI itself?

In this raw conversation, Kenny and Jake unpack the birth of LOOM—their flagship suite designed for persistent, AI-native code understanding. They reveal the core limitations of massive context windows in models like Claude, and how LOOM bridges that gap with game-changing features:

• The MCP Interface: 31 tools that let any MCP-speaking AI (Claude Code, Cursor, Windsurf, and more) truly understand a codebase—not just search it, but reason about functions, risks, blast radius, and dependencies.
• The Context Bridge (get_context()): A single call that compresses thousands of data points into ~500 smart tokens, tailored to intents like onboarding, editing, debugging, or PR reviews.
• The Living Document: LOOM's long-term memory that synthesizes architectural insights ("god functions," tight coupling, dead code hotspots) and gets smarter with every AI interaction—benefits compounding from the first scan to the tenth.

They get honest about the painful wall they hit: tons of powerful infrastructure (147,000 data points, massive graphs, caches) generating only a handful of insights—until the breakthrough that "intelligent" means deriving behavioral summaries, file zone maps, and patterns directly from code structure, no LLM required for the heavy lifting.

Looking ahead, they share their four priorities to supercharge the Living Document: function behavioral summaries, file zone compression, richer pattern extraction, and merge learning across sessions. The big vision? Point LOOM at any codebase, let it scan once, and empower any AI to grasp it deeply—without ever reading source code—answering "what does this do?", "what breaks if I change it?", and "what should I watch out for?"

This episode is a must-listen for developers, AI builders, founders navigating tech pivots, and anyone betting on the future of intelligent systems. Silicon Harbor isn't just consulting or coding—they're redefining how code lives, learns, and evolves in the AI era, right here in the Lowcountry.

Intro pre-roll

Outro post-roll