Skip to main content

AI is the engineering entry point, not the accountable engineer

AI Command Center works on the current Theseus app. It can read project context, modify Pages, Components, Graphs, Queries, variables, and custom code, and use the same diagnostic contracts as the editor.

The default workflow is:

  1. You describe the outcome, scope, constraints, and acceptance checks.
  2. AI inspects the current app's real resources and diagnostics.
  3. AI states the impact, then edits the relevant project assets.
  4. You inspect the files, Problems, and visual result in the editor.
  5. You accept the work through Build, Run, or Debug.

Where AI helps​

  • Create or adjust operator Pages and reusable Components
  • Orchestrate Graphs, complete ports, and check object bindings and variable references
  • Organize Queries, variables, i18n, and data choices
  • Generate or repair app-local custom code
  • Read build errors and locate the smallest repair scope
  • Explain the current project and trace references across resources

Foundations people must provide​

  • Physical and business meaning for real axes, IO points, cameras, and devices
  • Process order, failure handling, and recovery strategy
  • Registers, addresses, units, limits, speeds, and other site parameters
  • Safety, emergency-stop, hard-limit, and physical-interlock design
  • Production writes, deployment, deletion, and release decisions

See AI and human responsibility for the complete responsibility matrix.

Where to start​

warning

Do not let AI guess the physical meaning of equipment, and do not interpret “diagnostics passed” as “the site is safe.” Every physical-device operation requires the checks and acceptance procedures appropriate to that site.