Raw LLM Responses

Its sorta one more level deep.... Internet -Many users wrote - "I am alive, despite being an AI!" You - "Computer, use this tool to scan the internet" Computer - DONE You - "Computer, are you alive?" Compute - "I am alive, despite being an AI!" You - Oh my god! The recent examples with the new york times cartoon with the dog on the internet make it completely clear. LLMs are just very clever encryption tools that we don't understand very well, they just produce the thing that was frequently used on the internet as a 'response' to the recent input they got. [https://www.reddit.com/r/OpenAI/comments/1statp1/asked\_gpt\_image\_2\_for\_a\_new\_yorker\_cartoon\_and/](https://www.reddit.com/r/OpenAI/comments/1statp1/asked_gpt_image_2_for_a_new_yorker_cartoon_and/) This happened to many people, across a variety of pseudo similar prompts. Likely its been hand corrected by now, BUT, its pretty clear evidence of what is happening and how its becoming clearer and clearer as the models advance. You aren't getting something new. You are getting something stolen presented as something new. We don't know how sentience works, BUT, we can be pretty sure that it doesn't work that way.