Fuck AI

The Turing test is more showing how well a program can convince a person they are human. Clearly the past few years has shown that people can be easily fooled in many ways. The test was never meant as an indication of intelligence behind the messages, only whether or not they were good simulations of human discussion. The 2025 version of ChatGPT rated 73% on such tests, and I imagine it and other LLMs might do even better now with the extra tools they use to help context and memory. They still aren't "thinking". They're doing something though, and the usual "just a complex autocomplete" dismissal isn't recognizing that.