Blog
Insights on business intelligence, legacy systems, and AI-powered development.
-
From Production Exception to Fix in Minutes
A critical production error used to mean days of investigation. Now we trace, fix, test, and deploy in under 30 minutes. Here's the system that makes it possible.
-
Your AI Strategy May Depend on Models You Can't Control
Enterprise AI is being built on a fragile foundation. When frontier providers change their models, your workflows break. Open models change the economics of control.
-
The Trust Gap: Verification and Validation in a World of Probabilistic AI
When engineers write code, they write tests. When an LLM generates output, what's the equivalent? The answer will shape where AI can be safely deployed.
-
The 1-10x Reality: AI Productivity Is Earned, Not Promised
AI doesn't replace humans—it amplifies them. But that amplification isn't automatic. After 200,000+ sessions, here's what we've learned about trust, autonomy, and real productivity gains.
-
Why Your AI Makes Different Decisions Every Time
AI tools are inconsistent by design. Every session starts fresh, every decision is made in isolation. Here's how to build the consistency layer your business actually needs.
-
The Institutionalization Gap
Every AI agent session produces decisions, patterns, and context. Most of it never gets captured. Here's what we built to close the gap.
-
Your AI Agents Need Governance, Not Just Guardrails
AI agents are getting more autonomous, but not more aware. Guardrails stop bad behavior—but governance ensures consistent behavior. Here's the difference.
-
Why Off-the-Shelf AI Fails Growing Companies
ChatGPT knows everything—except your business. Here's why most AI pilots stall, and how to close the context gap.
-
The Two Things AI Assistants Keep Forgetting
AI coding assistants have no memory. We solved "what we decided" with persistent Decisions. But we forgot about "how we do things." Enter SOPs.
-
Building a Content System That Works for Introverts
Most advice for "building your personal brand" assumes you enjoy self-promotion. I don't. So I built a system that extracts content from work I'm already doing.
-
Every Story Dreams of Becoming a Test
Where do user stories go to die? Most teams have good capture OR good tracking. Rarely both. Here's how to close the loop.
-
The Gap Between Reactive and Proactive Security
Security tools tell you when something is wrong. But they don't tell you when something is about to go wrong. That gap is where the real risk lives.
-
Show, Don't Sell: The Art of the Invisible Pitch
The best way to introduce clients to something new isn't a pitch deck. It's using it in front of them until they ask about it.
-
Token $0: Where AI Pricing Is Headed
AI inference costs are racing toward zero. Not because they have to — because the giants want them to. The AI is the free refill. Your data is the Whopper.
-
Everyone Has GPT-5. So What?
GPT-5, Claude 4, Gemini Whatever. Your competitors have the same AI you do. The model isn't the moat. Context is.
-
The Best Architecture Is the One You Can Delete
We added MinIO for "proper" file storage. Fought CORS for an hour. Then asked: do we actually need this? Sometimes the best architecture is no architecture at all.
-
Your Data Has Answers. You Just Can't Ask It Questions.
Most businesses are drowning in data but starving for insight. The problem isn't the data—it's the translation layer between questions and answers.
-
The Problem With Dashboards
Dashboards show data. Executives need answers. There's a fundamental mismatch between what BI tools provide and what decision-makers actually need.
-
Why I'm Betting on "Mini-Palantir" for the Rest of Us
Enterprise companies have Palantir. Everyone else gets spreadsheets. The gap is enormous—and AI just made it possible to fill.
-
Graph Databases Are for Discovery, Not Documentation
Why your business intelligence tools can't answer "Why?"
-
The Fallback Trap: Why LLMs Keep Adding "Safe" Defaults That Aren't
When systems mask problems instead of surfacing them, the real cost shows up months later—usually during a board meeting.