Would you like to undergo a surreal, kind of boring, yet also kind of interesting philosophical experience? Then I invite you to try my newly patented “Chinese Room GPT”! That’s right, it’s time for you to play John Searle in this awful simulation of a simulation!
The Chinese Room is a thought experiment devised in 1980 by the philosopher John Searle. Now, before I dive in, here is a little overview of how this blog post is going to unfold.
First, I will explain the Chinese Room. Note that I will get into a bit of philosophy in this section, so if that is not interesting for you, after you read the summary of the original thought experiment, just go straight to the second section, “Simulating the Simulation”. There, I will discuss how a virtual Chinese Room could actually be built using ChatGPT 4 (or any LLM-based chatbot, as the baseline capabilities of all such chatbots is rising).
There are two versions of Chinese Room GPT I will discuss: the four-window and one-window version. The four-window version requires software, the one-window version just regular access to ChatGPT. Regrettably, I do not have the coding skills to build the four-window version, so I must leave that task to whoever might be interested enough to try to develop the necessary software. Yet, fear not! You can play the one-window version just by copy-pasting a single prompt (and you can also tinker with that prompt to your heart’s content to make it more interesting or functional).
Finally, I will close this blog post by showing you a sample of a Chinese Room GPT I am running right now (Tuesday, April 2, 2024), and in the comments section you can tell me which of the inputs you think I should use.
Continue reading “Chinese Room GPT”