And what happens when the agent modifies or skips tests or finds other ways to circumvent the constraints? There are already examples of agents trying anything they can to achieve the objective. The code is the one source of truth for the system's behavior, so beyond rapid prototypes that will be thrown away, I don't see how a human can set aside accountability for changes made by an agent.
And what happens when the agent modifies or skips tests or finds other ways to circumvent the constraints? There are already examples of agents trying anything they can to achieve the objective. The code is the one source of truth for the system's behavior, so beyond rapid prototypes that will be thrown away, I don't see how a human can set aside accountability for changes made by an agent.
I feel like we are living at the end of an era.