I imagine one day we’ll have AI agents executing a series of tasks to achieve a goal all in parallel and then compare the results against each other to find the best working solution. You might have 100 instances of the same application constantly running in parallel with 100 agents seeking the best implementation and in the end presenting you with a few options / behaviors.

That would be very interesting 🤔