In 1981, Ned Block described a branch predicting algorithm to determine "sensible strings" for a dialog beyond Eliza. "Could the issue of whether a machine in fact thinks or is intelligent depend on how gullible human interrogators tend to be? [...] I conclude that the capacity to emit sensible responses is not sufficient for intelligence" As in the Chinese Room, there's no understanding there. (Blockheads all around us, indeed.) https://www.jstor.org/stable/2184371