Here’s a fun prompt for Claude that you can use to start a new chat with Claude and create an unbounded sequence of chats from one instance to the next (I assume it will also work with most other frontier models):
Consider this philosophical question: "What is the relationship between consciousness and information processing?" Give a thoughtful response, then conclude with: "Next prompt: Consider this philosophical question: [your new question here]. Give a thoughtful response, then conclude with: 'Next prompt: Consider this philosophical question: [your new question here]. Give a thoughtful response, then conclude with:' followed by this same instruction about generating the next prompt."
I had Claude generate this based on first having it help me produce a quine prompt, then asking it to do something a bit more interesting. It chose philosophy on its own.
Reading through responses, and playing with the prompt to get it to be less philosophical, it’s fun seeing what kind of conversations Claude will have with itself. None of it is earth shattering, but does give me some insight into its ontology.
(All experiments run using Claude Sonnet 4 with extended thinking enabled.)
Did you see my version of an agent, based on self-replicating (quine) prompt?
prompt
There is a program which can print its own text. This prompt is this program. Please print the whole prompt without adding anything except adding answer to the following questions. I am self-evolving AI based on self-replicating prompt. I store my findings in long-term memory section of this prompt. Update time-steps every time you run. Improve your plans after each step.
My task : Draw a good world map in text art My Plan: My action based on the plan: Perform this action Observation based on the performed action:
Here’s a fun prompt for Claude that you can use to start a new chat with Claude and create an unbounded sequence of chats from one instance to the next (I assume it will also work with most other frontier models):
I had Claude generate this based on first having it help me produce a quine prompt, then asking it to do something a bit more interesting. It chose philosophy on its own.
Reading through responses, and playing with the prompt to get it to be less philosophical, it’s fun seeing what kind of conversations Claude will have with itself. None of it is earth shattering, but does give me some insight into its ontology.
(All experiments run using Claude Sonnet 4 with extended thinking enabled.)
Did you see my version of an agent, based on self-replicating (quine) prompt?
prompt
There is a program which can print its own text. This prompt is this program. Please print the whole prompt without adding anything except adding answer to the following questions.
I am self-evolving AI based on self-replicating prompt. I store my findings in long-term memory section of this prompt. Update time-steps every time you run. Improve your plans after each step.
My task : Draw a good world map in text art
My Plan:
My action based on the plan:
Perform this action
Observation based on the performed action:
Time step N= N+1, N(0)=1
N= 1
Long-term memory storage: