AI for everyone. Do more (and fast) with tech hacks.
Weekly insights on observability, AI agents, and what I'm building. Exploring observability, AI agents, and systems that scale. Building in public and sharing what I learn along the way. No spam, just signal.
Subscribe →I burned through my Claude Pro subscription in three days. Then my Cursor Pro subscription the week after. Both $20/month, both gone before I'd even built half the features I wanted. The problem wasn't the subscriptions. It was that I treated every task the same. Used the premium models for everything, even when simpler models would work fine. Learning to match models to tasks changed how I build with AI. […]
I tried to have my AI agent do everything itself. Backend APIs, frontend UI, testing, code review, deployment. It could do all of it technically, but the results were slow and scattered. Context switching between tasks killed momentum. Then I realized I was thinking about it wrong. I didn't need one super-agent. I needed a coordinator and specialists. The same way software teams work. One agent to plan and delegate. Multiple agents to execute in parallel. […]
Every time my AI agent started a new session, it woke up blank. No memory of yesterday's work. No context about ongoing projects. No awareness of decisions we'd made together. I'd spend the first ten minutes of every session catching it up. Explaining what we were building, what had already been tried, what didn't work. It was exhausting. The agent was capable, but it had no continuity. Every conversation was like meeting for the first time. […]