Adam Greenfield on Nostr: I don’t see how that branch of anti-AI sentiment that is primarily concerned with ...

I don’t see how that branch of anti-AI sentiment that is primarily concerned with minimizing its novelty – the “stochastic parrot” line – is tenable after recent events.

Before you reply with a lecture on how LLMs work, I’d ask you to consider that it is possible to understand how these systems function in some depth and sophistication, while also granting that a degree of emergent behavior was inevitable once they were deployed at scale.

And while we’re most used to concepts like “creativity” or “adaptive response” being used to characterize the behavior of subjects, I think we must also grant that actors without interiority, worlding capacity or a sense of self can do those things too. (Why should what is true of slime molds be any less true of sufficiently complex software agents?)

I’m not asking you to approve, and still less to endorse any description of the situation offered by commercial actors. I *am* asking that we contend with the reality of what we’re confronted with.