Are the machines’ goals equivalent, compatible, or incompatible?
In the first case, cooperation is no problem, the second case represents a rather unlikely scenario, and in the third case, you probably won’t see much cooperation.
The second case may be more plausible in a non-zero-sum environment, where the merged agent can achieve more of each prior agent’s goals than they would each achieve in a competitive setting. Think of the Superhappies’ propensity to merge goals with other species as an analogue.
Are the machines’ goals equivalent, compatible, or incompatible?
In the first case, cooperation is no problem, the second case represents a rather unlikely scenario, and in the third case, you probably won’t see much cooperation.
The second case may be more plausible in a non-zero-sum environment, where the merged agent can achieve more of each prior agent’s goals than they would each achieve in a competitive setting. Think of the Superhappies’ propensity to merge goals with other species as an analogue.
Companies merge sometimes—that seems like a kind of proof of concept. They are usually trying to gain mass in order to defeat third parties.