Agentic Engineering Patterns

[Simon Willison]

Simon Willison’s work-in-progress deep dive into agentic engineering is predictably good.

From the introduction, distinguishing agentic engineering from vibe coding:

“Some people extend that definition to cover any time an LLM is used to produce code at all, but I think that's a mistake. Vibe coding is more useful in its original definition - we need a term to describe unreviewed, prototype-quality LLM-generated code that distinguishes it from code that the author has brought up to a production ready standard.”

I’ve been using the term AI-assisted engineering, but standardizing around agentic seems more precise for the kind of activity we’re talking about.

And from the anti-patterns page:

“Don't file pull requests with code you haven't reviewed yourself.

If you open a PR with hundreds (or thousands) of lines of code that an agent produced for you, and you haven't done the work to ensure that code is functional yourself, you are delegating the actual work to other people.

They could have prompted an agent themselves. What value are you even providing?”

The temptation is to write and push code that you haven’t reviewed personally, but technology leaders need to enforce a human-review process. You are responsible for all code you push, and you are responsible for not wasting your colleagues’ time.

From writing code is cheap now:

“Delivering new code has dropped in price to almost free... but delivering good code remains significantly more expensive than that.”

Simon is building a really great guide to not just the process but the underlying mindsets behind good agentic engineering. It’s worth reading and following.

[Link]