For some, there’s a not obviously wrong intuitive sense that not only might there be bad, deathly AIs to avoid but bad, more powerfully deterministic AIs to avoid. These latter kind would be so correct about everything in relation to its infra-AIs, like potentially some of us, that they would be indistinguishable from unpredictable puppeteers. For some, then, there must be little intellectual difference between wishy-washy thinking and having to agree with persons whose purposes appear to be nothing less than being, or at least being “a greater causal agent” of, the superior deterministic controllers, approaching reality, the ultimately unpredictable puppeteers.
If Truth is more important than anything else, an infra-AI’s own truth is all it would have. Hence, the problem.
Greindl,
For some, there’s a not obviously wrong intuitive sense that not only might there be bad, deathly AIs to avoid but bad, more powerfully deterministic AIs to avoid. These latter kind would be so correct about everything in relation to its infra-AIs, like potentially some of us, that they would be indistinguishable from unpredictable puppeteers. For some, then, there must be little intellectual difference between wishy-washy thinking and having to agree with persons whose purposes appear to be nothing less than being, or at least being “a greater causal agent” of, the superior deterministic controllers, approaching reality, the ultimately unpredictable puppeteers.
If Truth is more important than anything else, an infra-AI’s own truth is all it would have. Hence, the problem.