Perhaps related to this mother goose nursery counting rhyme?
One, two, three, four, five,
Once I caught a fish alive.
Six, seven, eight, nine, ten,
Then I let it go again.
Why did you let it go,
Because he bit my finger so!
Which finger did it bite?
This little finger on my right!
https://en.wikipedia.org/wiki/One,_Two,_Three,_Four,_Five
Perhaps in a bizarre twist of fate, GPT learns similarly to many young humans or even adult learners on new languages: by using nursery rhymes.
Edit: to add, I wonder if the is used as a vector/index of relative indirection. That would mean clusters of meaning around whatever level of the is being used.
In the end all language could be a 1 op instruction just compressed with a distance function from the root position. Almost like a MOV based cpu—perhaps transport mov based with some ‘aliasing’ (mov pointing to N long mov). Also to be maximum efficient for compressibility and durability there would seem to exist something like this, as it appears this is what genes do.
It seems to me that in 2020 the world was changed relatively quickly. How many events in history was able to shift every mind on the planet within 3 months? If it only takes 3 months to occupy the majority of focus then you have a bounds for what a Super Intelligent Agent may plan for.
What is more concerning and also interesting is that such an intelligence can make something appear to be for X but it’s really planning for Y. So misdirection and ulterior motive is baked into this theory gaming. Unfortunately this can lead to a very schizophrenic inspection of every scenario as if strategically there is intention to trigger infinite regress on scrutiny.
When we’re dealing with these Hyperobjects/Avatars/Memes we can’t be certain that we understand the motive.
Given that we can’t understand the motive of any external meme, perhaps the only right path is to generate your own and propagate that solely?