Automating the boring stuff

I've spent the last few weeks thinking about what AI is actually good for, and it's not what the marketing hype suggests. It's not about replacing developers or making grand predictions about the future.

It's about the boring stuff.

You know those tasks that make you want to scream? The repetitive config files, the boilerplate, the data transformation pipelines that take 30 minutes to write but are intellectually trivial. That's where AI actually shines.

The Pattern I've Found

Here's what works for me:

1. Identify the tedious part — not the creative part. Creative work still requires you.
2. Give AI clear instructions — be specific about format, style, and constraints.
3. Review and iterate — treat the output as a starting point, not gospel.
4. Build small — automate one thing at a time, not entire workflows.

Real Examples

• Database migrations: I describe the schema I want, AI generates the SQL, I verify it.
• Test data generation: Instead of manually creating fixtures, I write a prompt and tweak the output.
• Documentation: Tedious but necessary. AI handles the drafting, I add the personality.

The key is knowing when to use it. When a task requires judgment, taste, or creativity — that's you. When it's repetitive, mechanical, and verifiable — let AI handle it.

The goal isn't to work less. It's to work on the interesting bits.