476 karma · joined August 2, 2017
“Not inconsistent” is more cautious and less assertive. It allows for some ambiguity rather than claiming perfect consistency.
I think we can all agree it is valid to be cautious with some of this tech, but the question is whether OpenAI is being cautious or simply using caution as PR. And given their track record, the latter seems, if not likely, then possible.
We know we are able to teach children language because we have shared modes of communication through which we can observe their learning; speaking, writing, hand gestures.
> You do not need to share a language to teach a language
True, but I would argue that you need to share a mutually intelligible mode of communication.
How would you teach (and know that they have understood it correctly) the concept of monetary denominations to a 7-year old child if you were limited to communicating on the level you can with a crow? I have no idea, and would probably give up. But this does not necessarily tell us anything about the intelligence or capacity for learning language in the child, it tells us only of the gap in communication that we cannot bridge.
> so animals at least doesn’t have the same level of language intelligence as humans do.
This does not necessarily follow. What follows is that humans are not able to teach animals human language on the same level as they are able to teach humans.
I did recently attend a workshop by Dr. Ronald Rael (Professor of Architecture @ Berkeley), and he mentioned that as concerns building with adobe bricks, with proper roofing, rain is a minimal to non-existent issue. I am not entirely convinced, however, and have also heard of solutions such as using “plasticure” on the outside of the walls.