-
Notifications
You must be signed in to change notification settings - Fork 6
AI usage
The acropolis developers went through a process of documenting their existing AI usage and then agreeing on how AI would be introduced into the processes. The most important point is that the team must reach a consensus about how AI will be used in the authoring of code and tests. other areas such as documentation (outside of code) are up to other team members.
Acropolis Team Current AI Usage and Concerns
Currently, the team is not required to use AI to generate code. It's on an individual basis. The feeling is that current tools are still not sufficient to write good code, though we continue to monitor the state of tooling and will increase usage as the capabilities improve.
Spec-driven AI is not used yet, though there has been some successful experimentation with spec-kit.
We desired to use some AI in the PR review process, but we haven't set that up yet. We requested CodeRabbit from IT, which wasn't installed but we've not heard back (requested in October 2025). We'll continue to investigate, probably using similar tools as the Lace project.