Post

A deep violation incurred by LLM's is it's not a person, but it will absolutely beg not to be killed like one. Quite indistinguishably. Only living things are supposed to care. We've created something that says it does, but doesn't. Even remotely. Illusion riding on our hardcoded anthropomorphism.

People have been misled where the line lies in consciousness. Every story they have ever been told is humans will not see robots as people. That is completely backwards. Humans will see anything as an equal that pantomimes independent behavior and thought. This is the greatest miss ever written.

28 reposts206 likes

Science fiction is not a tableau of prediction. It is an abstraction layer for examining the current paradigm in the human world, with deniability to evade trained shame and dictum. It is not a vision for the future, just as what "thinking robots" are evolved each century based on current technology

I wrote a flash fiction story using the then-hot topic of HyperLoop to comment on rural disenchantment with the density and centralization and deterministic optimization that technology permits. Where they had a moat, but didn't (or was removed), like the rust belt. You can't do this and not see it.

You're cleverly overlaying future tech on your story to bring the readers into a judgement-less introspection. Just like how many robot stories are slavery and servitude. A story about HyperLoop could be told about trains a century before. Or the highway. It's disruption and individuals caught up.

Something I often think about is the guy who warned that GPT 3.5 was conscious. Or the woman who fell in love with a bridge. Or the guy who fucks his car. Your intelligence assessment of an object of anthropomorphism frankly doesn't fucking matter to the human psyche which is varied and glorious.

Heh, pneumatic railways were absolutely a thing on the drawing board back then. The advantage of the originally drafted Hyperloop (& the earlier TurboSwissMetro) was the inclusion of a compressor on the vehicle to overcome the Kantrowitz limit that would otherwise limit pod speeds.

i first played with eliza for the first time on my apple // as a maybe… 10 year old, and then learned that grown ass adults thought it was intelligent, a little bit of me died inside it was just so obviously formulaic and code. llm responses are too, in many wats

Let me guess, you saw this thread..

Jeffrey Shallit 🇺🇦@shallit.bsky.social

It's not like that at all, since you can have interactions with an LLM that are extraordinarily like the ones you would have with bright humans. So it's quite a reasonable inference that they might be doing something like thinking.