I don't think we should take it as a given that these are truly precursors for AGI.
We may be going about it the wrong way entirely and need to backtrack and find a wholly new architecture, in which case current capabilities would predate AGI but not be precursors.
I call them precursors because we would anticipate an ASI to be able to do thsse things. Perhaps necessary conditions would be a more appropriate term here.
We may be going about it the wrong way entirely and need to backtrack and find a wholly new architecture, in which case current capabilities would predate AGI but not be precursors.