Kurzweil’s 2029 Turing-test wager

The rules are the story. A forecast that uses a named test should be evaluated against that test, including who judges and what alternatives they compare. Moving the goalposts to “a chatbot fooled someone” would answer a different question.

This entry is included among capability forecasts, because passing a test is neither a prediction of catastrophe nor an automatic proof of every possible meaning of AGI.