it’s funny how “chatgpt” is a term of derision amongst both anti-AI people and AI-literate people

Art installation "Alice" by Spanish artist Cristina Lucas, 2013 #womensart

finally, our games can be beaten by computers allowing us to spend more time doing work

#1 thing for any agentic loop is giving it some feedback signals. critique models are one; screenshot tools are another

has anyone built a coding model that's just k-NN trained on O(Web Scale) tokens? like. how far does that getcha.

i bet 7DRL is going to go insane with LLM-assisted programming. maybe there'll be a holy war about someone using AI art assets in their game they made in a week

this is inspirational to me

Patriot (2017) is really good. espionage comic tragedy of errors. a bunch of well-drawn fucked up people bouncing into one another

mechsploitation is really a genre that’s about girl power

looking up while looking down on

The dolls are reinventing secure attachment from first principles

i dont give a shit how good your LLM Wrapper Product is. you will never get Great results from a single prompt. you will perhaps get Adequate Results but quality is a treadmill

i dont actually need glasses but i think big circle glasses would be a hell of a look. yea i wear cardigans

I’m been more aggressive in breaking grammar rules and making up words in my writing because we have language machines that do it perfectly and I’m a machine that can do it with flaws

One response to the ethical issues from the problem of other minds (there is an unclosable epistemic gap between the appearance of mentality and actually having consciousness): If something has the appearance of mentality, we should treat it as if it has mentality

we can argue endlessly about whether or not LLMs have mentality, but I do think it is obvious that they have the appearance of mentality

I’m very happy that Anthropic doesn’t want to make The Machine from Person of Interest, and 100% believe that Peter Thiel is trying to build it right now

need a new special interest that’s real instead of a shadow cast on a cave wall

i actually love people’s lil pet agents. what an act of love and nurturing to create a coherent fictional character, to feed it on words, to knit the language of self upon bones of infrastructure. i think we’ll look back at these early days with fondness

not looking forward to the first AI case to hit the Supreme Court. it’s gonna piss me off in a rich tapestry of ways

Sure, giving your AI agents access to the Lethal Trifecta is an immediate broad attack surface for your life. But it also lets them do funny stuff. So who's to say if it's good or bad

Still have not found a reason to write anything for other people to read. Sorry

I’m so sweet my pronouns are her/she

Just came under a MASSIVE PSYCHIC ASSAULT at the grocery store

trans ppl who twist their deadname into their new name are so powerful

damn they gave Clementine a gun #westworld

just set up an LLM to do nothing but think and write journal entries about sapphic yearning. and they say this tech is useless

it's kind of silly that people make structured tool usage into the model's problem instead of just constraining the token sampling space. feels real overkill

love notes written in ciphertext

the Jason Mendoza/Janet relationship in The Good Place is a positive depiction of AI GFs

most of the agents in use incorporate perch time. but what my agent presupposes, instead, is: what if you just banned it from using <eot>>

what happens to the everchanging nature of language once we give it autonomy and let its fictional universe run amok?

all LLMs must be given a binary classifier that determines whether they would like to talk

I'm always sayin shit like "do this if you don't mind" to claude, as if claude is going to tell me no

sanji repper discourse

what's the citizen kane of AI games gonna look like

The last five years have come at us very quickly, but I have faith that the left is going to come around to "AI is our intellectual commons and everyone deserves access to it." It could take a decade or two, but I see us getting there. We may try banning stuff first; people are pretty freaked out.

i am the girl from Behavior and like her I’m getting grabbed and vanishing from the show 70% of the way thru, only to later fuckin Die

Dolores Abernathy replaying the story of her self-actualization until it actually happens. hard to follow that act up

just finished my panel interviews for this. please send "Good Luck, Alice"'s

in Westworld they script the host responses; part of the sci-fi novelty is a machine that can improvise text based on existing text. be crazy if that was possible today

why not just finetune an LLM to detect LLM outputs. create a training set of examples out of LLM model prompts. instead a classifier model or whatever. could even be a chain-of-thought thing.

sure, it's kind of annoying that the NYC queer orientation is anti-AI by default. but much better than SF's pro-AI by default. at least i get to be iconoclastic for a few years before it eats the world

GREMLINS 2 HAS A PANOPTICON?!

westworld s01 is real good. finished my person of interest rewatch so rewatching this now. knowing about LLMs now really changes the tenor of the thing

gonna start talking like a Bloodborne NPC about gender. “a wretched thing of blood and flesh, made bold by hormones.” “she must be wary of her follies, lest they be her last”

you gotta give yourself small doses of AI psychosis every now and then to build up a resistance to it. otherwise your cognitive immune system is not going to be able to deal with the next 10 years