You just missed the hint

Pas capté, c’est ton chat et c’est de l’AI?

Also, LLM support is shipping with Dagger 0.17, we'll share more of this soon!

Check this out: this Pull Request (github.com/samalba/demo...) was written by a model—it scanned the repo, optimized the Dockerfile, and opened a PR on its own. Then another AI agent jumped in and reviewed the changes (the second comment). It’s open if you’re curious: github.com/samalba/agents.

My take away from the article is that an "agent" is a long-running workflow.

Sure, like containers, frameworks, testing libraries, etc... It's all software indeed. The term "agent" is confusing though, "AI workflow" is more accurate to my developer mind.

I've gone the dagger.io route for project specific tooling and generally containerized setups, for both building and as a containerized task engine

Haha! TIL I have a signature laugh! I love it 🤣