Furthermore, it is absolutely unclear whether it is even possible to develop AGIs.
Yes, I think we need "modified turing test". Sure, it can be difficult or impossible to tell the difference between a LLM and human, however, that LLM was seeded with conversation from conscious / intelligent humans, thus it renders the test invalid.
However all this is a moot point. Humans (esp since giving up God), love worshiping the various idols they create....therefore its conceivable a huge part of humanity decides LLMs are in fact "intelligent" and even "conscious" and treats them accordingly.
I will try to find any prompts to let gpt-4o reveal that its just code or even sending it in a loop
reply