Blog

A Decentralized Architecture for Visual Forensics Introduction   Swarm-based image detection offers a powerful new approach to visual forensics. Instead of relying on a single model to assess whether an image is real, tampered, or fabricated, this architecture distributes the work across many lightweight, specialized agents. Each agent analyzes only part of the image or focuses on a specific kind of feature, allowing the system to detect subtle inconsistencies and

The Nobel was dead. Shocking news — but not as shocking as the image he held in his hand: a still frame, grainy and damning. It showed the Protectorate — chief advisor to the royal — mid-altercation, the crack of a neutron pistol frozen in time. Enter the investigator. She didn’t look like a detective. More like a seer, or one of those wandering subjects paid to uncover lies. She

Case Study B: Revisiting the Peanuts and Apples Problem   In a Substack article I described another precursor swarm, this problem is often sometimes thought of as a Turing test for LLMs but can be solved with multi-agent technology. Layers and Peanuts Another common “unsolvable” problem may actually be solvable with advanced LLMs that employ agents. I explored how this may be solved in depth in a previous article, but

Case Study A: Live Legal Tech Application with Emergent Swarm Principles   In a recent legal technology engagement, the author designed and deployed a production-grade application that implemented a precursor to the swarm agent model. The client—a firm focused on streamlining court filings, waiver forms, and appeals workflows—required an adaptive system capable of operating across fragmented jurisdictions and highly variable procedural rules. The solution employed modular task agents loosely coupled