Blog
-
Why I'm Building One AI System Instead of Five
I run several products with different problems, but I keep coming back to the same conclusion: the AI layer should be one system, not five bolted-on copies.
-
The Self-Improvement Loop Is Mostly Just Good Code Review
I want Nova's agents to review and refactor Nova's own code. The hard part isn't the agent — it's building enough verification scaffolding that letting one loose is a reasonable idea.
-
The Agents Were the Easy Part
Nova is now in its production orchestration phase with 34 agents running. The lesson so far: writing agents is cheap, and deciding which one should be talking is where all the real engineering lives.
-
34 Agents Later: What I've Learned About Building a Multi-Agent System That Doesn't Fall Over
Nova has grown into 34 agents spanning chat, ERP work, and pipeline execution — here's what actually mattered as it scaled, and what I got wrong.
-
Why I'm Building One Dashboard Instead of Five Dashboards
I run four active projects and kept context-switching between four different mental models of 'what's going on.' So I'm building a single orchestration view instead of living in five tabs.
-
34 Agents Is the Easy Part: Nova Hits Production Orchestration
Nova now runs 34 agents across chat, pipelines, ERP specialists and language tooling. Getting them to exist was straightforward. Getting them to work together without me babysitting is the actual engineering problem.
-
34 Agents Is the Easy Part
Nova has 34 agents and just moved into its production orchestration phase. Spawning agents turned out to be trivial; deciding which one should answer, and what it's allowed to remember, is the actual work.
-
34 Agents in Production: What Running a Real Multi-Agent System Actually Looks Like
Nova, my personal AI system, just crossed into full production orchestration with 34 active agents. Here's what that actually means — and what nobody tells you about running multi-agent systems outside of demo conditions.
-
Thirty-Four Agents and a Question of Control
Nova, my personal AI system, now runs 34 agents across general chat, pipeline execution, ERP work, and more. Here's what 'production orchestration' actually means when you're the one who built it.
-
What Running 34 AI Agents in Production Actually Looks Like
Nova, my personal AI system, just hit the production orchestration phase with 34 agents spanning chat, pipelines, ERP, and language tasks. Here's what that milestone honestly means — and what it doesn't.
- Older posts