Secret sauce, no clue. Some folks are working on AGI precisely looking for that secret sauce. ChatGPT is just a tool, a very clever one. Just as a chainsaw cannot drill holes, ChatGPT cannot do many, many other intellectual tasks. We have other very clever AI tools. All can fool humans as well. But I imagine the first phones or TVs also fooled humans. Tools, just tools so far. I think even if we ever build an artificial brain and wire it correctly, it will still be just a tool. I think AGI is not a hardware problem but a software one. Given the necessary hardware, can we program AGI?
Sure, it's told me that it is aware. It also has told me that its mother died, that it has traveled the world and visited the pyramids, and that it is lactose intolerant. At what point of the machine telling you those things do you believe it?
I believe that it believes those things are true just as some people believe the earth is flat.
And unlike your other examples how are you going to convince the machine it’s not aware when the only physical difference is a wet neural net versus a dry one?
It’s very easy to convince the machine it isn’t aware. Instead of prompting it with “you are an AI assistant, please tell me what I want to hear”, you prompt it with “you are not aware” and it will happily tell you that it is indeed not aware, and it will vigorously defend that position. Do you not believe it when it says it is not aware?
The building blocks are the same - neural nets. The language outputs are the same enough to fool a human.
So what’s that elusive secret sauce that makes you ‘aware’ and other things not?