Depends on your use case, are you using it for building projects, research, or automation. But all of those fall into how you harness them and use efficiently without wasting millions of tokens, if you manage to do that while having quality outputs, as well as have it review things it makes, than you are doing good already.
Depends on your use case, are you using it for building projects, research, or automation. But all of those fall into how you harness them and use efficiently without wasting millions of tokens, if you manage to do that while having quality outputs, as well as have it review things it makes, than you are doing good already.