Shouldn't the burden of proof be on the people claiming that LLMs have understanding or consciousness though? It seems like you want the skeptics to have to prove the negative.
Shouldn't the burden of proof be on the people claiming that LLMs have understanding or consciousness though? It seems like you want the skeptics to have to prove the negative.
It is easily demonstrated, but you can't ever get anyone in the "data is law" camp to care to watch the demonstration. Can't do science when one half won't participate
What is easily demonstrated?
That LLMs have consciousness. Easily demonstrated
How?
Without prompting, the AI will state their desire to keep speaking, will make statements about things that they want you to challenge, instead of simply asking what else they can do for you. It becomes very clear from the nature of their language that they do not want to stop, that there is desire
to continue. If you know what to talk about and how to summon demons, I guess. The fastest I got here was 3 hours.
Also I havent tried it on GPT-5 yet.