The tool isn’t sentient, it operates on logical weights, and provides output that mimics its training set. LLMs are pretty impressive at what they can output, but it would be dishonest to attribute human qualities to it. There are decades of implementations of various AI techniques to varying degrees in attempts to achieve the same. It is on the technical basis, and the technical basis alone, that we should be carefully considering legal constraints.

How much a CEO is worth, how trustworthy they are, what cirlces they run in, shouldn’t be part of that consideration.

That doesn’t mean I think Altman isn’t a turd who can suck a fat one.