← All series
Working with AI in 2026
A practical series on context, iteration, prompts, reusable workflows, and human judgment in AI-assisted work.
- Part 1The Latent Space: The mental model that changed how I work with AIMost AI failures aren't the model's fault. They're navigation errors -- reaching into the latent space without knowing what you're doing.
- Part 2Context is the InterfaceThe context window isn't a log of your conversation -- it's the environment you're working in. How you build it, manage it, and know when to reset it is a learnable skill.
- Part 3Iteration as MethodThe first output isn't the work. It's the starting point. What changes when you treat AI-assisted work as a cycle -- generate, evaluate, refine -- rather than a prompt waiting for the right answer.
- Part 4The Prompt is a DraftMost people treat prompts as instructions. They're not -- they're the starting position for a context window that already has shape. Treating the prompt as a draft you revise before sending is the simplest way to change what comes back.
- Part 5Systems That Build SystemsThe most useful thing AI changes isn't individual tasks -- it's the economics of building the systems that do the tasks. When scaffolding gets cheap enough, you build infrastructure you'd have skipped. That infrastructure compounds.
- Part 6The IT Guy Who ListensAI changes the cost of execution. It doesn't change the cost of judgment. The person who keeps the diagnostic loop open, asks the question that reframes the problem, and knows what the system is actually for -- that's the human moat.