I agree that modern AIs would pass the Eliza test as you define it, particularly if given the right prompt. What is the value of the test though? It seems like it would be saturated as a benchmark, and also take a long time to administer.
I agree it doesn’t have much practical value. I thought it was “philosophically” interesting. This is because it seems an important benchmark in terms of our relationship with machines.
The hypothesis is that our intuitive / system 1 aporoximation to treating machines as thinking beings depends on a benchmark we already passed.
I wrote it down because i want to refer it in future pieces (like the one on ai and loneliness)
I agree that modern AIs would pass the Eliza test as you define it, particularly if given the right prompt. What is the value of the test though? It seems like it would be saturated as a benchmark, and also take a long time to administer.
Sorry for the late reply.
I agree it doesn’t have much practical value. I thought it was “philosophically” interesting. This is because it seems an important benchmark in terms of our relationship with machines.
The hypothesis is that our intuitive / system 1 aporoximation to treating machines as thinking beings depends on a benchmark we already passed.
I wrote it down because i want to refer it in future pieces (like the one on ai and loneliness)