Building Software That Lasts
What separates vibe coding from agentic software development? Getting a feature working is only part of the job when the software needs to remain maintainable, secure, and reliable as it changes.
In my latest video I share two examples from my own experience: using Claude Code to rework my WordPress site, and using a structured agent workflow to develop a software feature in a commercial software app stand-in.
We’ll explore architecture, context limits, automated testing, and how multiple agents can plan, implement, review, and audit a change—and why software development experience still matters.
- Why AI is changing software development
- Vibe coding vs agentic software development
- My WordPress site: a vibe coding example
- Asking Claude Code to rework the site
- What makes this vibe coding?
- Commercial software, architecture, and context limits
- Why AI changed my view of unit testing
- Planning work for multiple agents
- Inside an agent workflow
- Building software that stays maintainable
- Learning agentic development: what comes next