Posts
Notes on software, system design, and building with AI.
AI is an amplifier, not an equalizer
Every company has the same coding agents. The output is not the same. AI multiplies what you already bring, and it multiplies the gap with it.
The brain gets better at what you actually do
Neuroplasticity is real, but it is narrow and it fades. What that means for how I learn new tools, and for what I let an agent do for me.
Your agent should never see the key
Credentials belong outside the sandbox. I compare five agent runtimes and the gap between containment and delegated authority.
Agents are a CPU workload now
AMD and NVIDIA both announced racks for running agents, not models. The GPU runs the model call. Everything around it is CPU work.
