I’m a maker who happens to use software.

Coding agents let me build at a much higher pace than traditional development alone. I can explore more approaches, turn rough questions into working prototypes, and spend less time waiting for an idea to become concrete enough to evaluate.

That does not make engineering judgment less important. I still own the problem definition, architecture, tradeoffs, validation, and behavior of what ships. AI makes the implementation loop faster; it does not decide what is worth making or whether the result is trustworthy.

Faster code generation exposes gaps that slower implementation used to hide. Unclear requirements multiply faster. Weak review practices miss more. Testing, security, integration, deployment, observability, and operational ownership become the limiting factors—not how quickly someone can type the next feature.

I’m passionate about closing those gaps: better specifications, replaceable boundaries, automated verification, meaningful evaluation, explicit failure paths, and delivery systems that can keep up without lowering the standard for what reaches users.