Signalbox: Structural Trust for AI Development Workflows
Signalbox rebuilds the TRIP development loop as Emergent event topologies. Its gates, validators, and barriers make workflow trust a property of structure instead of prompts.
Writing on systems design, constraints, and developer experience.
Observations on how systems coordinate, why constraints shape behavior, and what makes APIs feel right.
How to Tell Whether Source Code Was Copied
Resemblance is the weakest place a copying case can start. What separates real copying from coincidence—and how generative AI changes the analysis.
The Pattern That Came Back: Adaptive Object Model in an Agent-Iterated World
A 1998 pattern called Adaptive Object Model was overengineering when humans owned the schema. Agents iterate too fast for that. Here is what stops SchemaForge from becoming stringly-typed chaos.
Accounting for Agents
Most accounting software was designed before AI could do the bookkeeping. That assumption is now a bug.
What Happens When You Constrain an Event-Driven System to Three Primitives
Constrain each component to one of three roles (Source, Handler, Sink) and complex coordinated behavior emerges without orchestrators, brokers, or coordination protocols.
Stable Agent Identity Without Centralized Control
Why agent systems break when you change URLs, and how separating names from addresses solves it.
When Users Route Around Your Constraints
What Roblox chat restrictions teach about designing constraints that guide rather than block.
Treating UI Regions as Independent Actors Makes Terminal State Manageable
How borrowing distributed systems patterns from Erlang creates maintainable, efficient terminal applications with independently managed regions.
Why Multi-Agent Systems Don't Need Managers: Lessons from Ant Colonies
What happens when you stop designing coordination and start constraining it? A research exploration of stigmergic coordination for LLM agents.
Designing APIs for the Pit of Success
A case study in how a programming language's type system can guide developers toward correct usage, and why that matters for security libraries.
Constraints as Design
The design philosophy behind acton-reactive: how intentional API constraints eliminate entire error classes at compile time.