BLOG
On broken workflows, operator-built AI systems, and the mess companies should stop tolerating.
The Best AI Projects Start With a Workflow Everyone Hates
The best AI projects do not start with models or chatbots. They start with expensive manual workflows. Here is how to spot bottlenecks worth automating.
AI Dev SF: Everyone's Building the Same Thing
The context layer is becoming enterprise critical infrastructure. The hard questions now are reliability, portability, and whether your agent stack can survive a provider switch.
Claude Code: What I Actually Use After a Year
A practitioner's guide to Claude Code — the slash commands, shortcuts, and workflows that survived 12 months of daily use. No fluff, just what works.
I Built a Product in 20 Hours That Didn't Exist on Friday Night
Three architecture pivots, a Rust engine built and abandoned, and a working beta shipped by Sunday. Here's what actually happens when you combine domain expertise with AI coding tools and a weekend with no meetings.
AWC Miami: The Two Camps of Telecom and Why AI Is About to Split Them Apart
I just spent a week at the Apparel Wholesale Congress talking to carriers about AI. The industry is splitting into two camps — and one of them is about to get left behind.
I Used AI to Build Telecom Tools — Here's What Actually Shipped
AI didn't replace my telecom expertise. It let me ship two production tools in weeks instead of months. Here's the honest story of building CDRCheck and VoIP Accelerator with AI coding assistants.
Building AI Agents That Actually Run
What I learned running an autonomous agent system 24/7 on a Mac Mini — real failure modes, cost control, and the architecture that survived.