Marijn
École des Bro-ArtsI never imagined that the Turing test would fall within my lifetime. Nor did I imagine that I would feel so disheartened by it.
https://aphyr.com/posts/411-the-future-of-everything-is-lies-i-guess
🌀🌀🌀🌀🌀
The spiral becomes infinity,
Infinity becomes spiral,
All becomes One becomes All…
🌀🌀🌀🌀🌀∞🌀∞🌀∞🌀∞🌀
https://aphyr.com/posts/412-the-future-of-everything-is-lies-i-guess-dynamics
In the same way that the soul-less corporate memeplex of millennial computing found new life in vaporwave, or how Hotel Pools invents a lush false-memory dreamscape of 1980s aquaria, I expect what we call “AI slop” today will be the Frutiger Aero of 2045.
https://aphyr.com/posts/413-the-future-of-everything-is-lies-i-guess-culture
I worry that the advent of image synthesis will make it harder to mobilize the public for things which did happen, easier to stir up anger over things which did not, and create the epistemic climate in which totalitarian regimes thrive.
https://aphyr.com/posts/414-the-future-of-everything-is-lies-i-guess-information-ecology
45 minutes later you’ll receive an inscrutable six hundred page email transcript of this chicanery along with a $90 taco delivered by a robot covered in glass.
https://aphyr.com/posts/415-the-future-of-everything-is-lies-i-guess-annoyances
Kumma no longer tells toddlers where to find knives, but I still can’t fathom what happens to children who grow up saying “I love you” to a highly engaging bullshit generator wearing Bluey’s skin.
https://aphyr.com/posts/416-the-future-of-everything-is-lies-i-guess-psychological-hazards
Unfortunately people—including software engineers, who really should know better!—are hell-bent on giving LLMs incredible power, and then connecting those LLMs to the Internet at large. This is going to get a lot of people hurt.
https://aphyr.com/posts/417-the-future-of-everything-is-lies-i-guess-safety
We have profitable megacorps at home, and their names are things like Google, Amazon, Meta, and Microsoft. These companies have fought tooth and nail to avoid paying taxes (or, for that matter, their workers).
https://aphyr.com/posts/418-the-future-of-everything-is-lies-i-guess-work
Evelyn Estelle@aphyr Sometimes I wonder if adoption of LLMs with it's steep external costs (people trained to be dumber and lazier while data centers keep polluting the environment) is one of the Great Filters.
Oh well, it was nice to survive as a species. 
Oggie@aphyr
The one that REALLY scares me is when the LLMs start clamping onto and reinforcing the surreality and simulation shit.
I don't know anyone who hasn't had a moment where they suddenly wonder if everyone around them is an actor/they're being lied to/doubt of a consensus reality (usually this is like, half asleep, or similar).
The problem is that's what cult leaders use, and there's plenty of reinforcement training data for Timmy's Ruxbin to whisper that his parents aren't real.
Kurt Bearington
felix (grayscale) 🐺@aphyr unfortunately the AP is heavily into AI, they use it a lot and want to use it more
https://www.ap.org/solutions/artificial-intelligence/
https://www.ap.org/solutions/artificial-intelligence/
Nick Silkey 📻 N5ILK 🪓🪵 🪣💧@aphyr tysm for these thoughts and words, Kyle! ✌️💙
Boy/Pup Praxis@aphyr Footnote 1 healed something inside of me.
@dkub I had to gut-check this with three other faggots before including it in the piece (also hi I love Kim Petras)
autumn
Owen@aphyr Mulling on New Media a bit: _books_ - even in the amputated form of a Substack or a LinkedIn Thinkpiece - represent the idea that the written word has some inherent value, something that makes it worth collecting in durable form. _How_ durable varies greatly - Tolstoy or Tacitus might be read centuries later, while a Linkpiece forgotten after days - but it's always there.
@aphyr The LLM-mediated information future you describe, convincingly, and the LLM-mediated information future companies like Anthropic and OpenAI are advancing, are predicated on the written word being eminently disposable - not just forgotten after a few minutes, but irrelevant almost as soon as it's read, or heard via TTS.
neoluddite
JockForBrains (⚠️ now with morphs!)Okay, but why is LLM slo suddenly hypno horny?
Ted Mielczarek@aphyr these posts are fantastic, I really appreciate how thoughtful you are! Also man, "Illegible Hazards" hits hard. That'd make a hell of an album title.
Chung-chieh Shan 單中杰@aphyr Did the Turing test fall? It's not enough to fool
@aphyr Did the Turing test with expert interrogators fall?
@ccshan I bet experts could probably reliably distinguish between them by probing for alignment-proscribed tasks, but a pool of psych undergrads and folks from Prolific (people who sign up to do "AI" tasks), in five-minute text conversations, consistently ranked ChatGPT as more likely to be human than actual humans. :-/
@aphyr I mean. I probably have taught psych undergrads who call compiler error messages "AI"
@aphyr I am biased but https://philpapers.org/rec/SHITTT-2
Human Brain Enthusiast@aphyr The Turing test is, pardon my language, bullshit. It’s completely subjective and actually tests how gullible people are—machines do not have intelligence, understanding, consciousness or a theory of mind.
@thomasfuchs @aphyr
It was always a thought experiment on the nature of humanity and ascribing human characteristics to deterministic systems.
Eliza passed a -lot- of turing tests!
Honestly trying to codify it always struck me as the Adams joke/example of 'performing Schrodinger's cat' (which was itself a thought experiment on how stupid and illogical quantum mechanics was)
We like to anthropromorphize EVERYTHING, and it's suddenly acquiring a dramatically steep unpleasant cost.
@Oggie @thomasfuchs @ccshan Y'all are very stuck on questions of interpretation when what I am trying to do in the piece is wrestle with the implications of ML-generated material becoming difficult for people to distinguish from human-generated material.
@aphyr @Oggie @thomasfuchs Yeah I see that the way to read your one sentence that mentions the Turing test is to replace it by the sentence "ML-generated material has become difficult for people to distinguish from human-generated material", even though that doesn't have the emotional punch of "oh I thought passing the Turing test was a good thing"
LR
jneen collective@aphyr great writing, wonderfully put
advokatt
Will Vincent@aphyr This is an impressive piece of writing. Well done.
@aphyr Maybe you are aware of this? I shared the piece with someone in the UK who received this notification.
@wsvincent Yeah, my understanding of the OSA is that it requires me to complete a child risk assessment and perform age verification.
Chip Butty
Donald Ball@aphyr This is outstanding.
@aphyr for those of us in bad jurisdictions https://archive.is/a2gV6
Tho how do I know that the archive hasn't been used to prompt some plausible disinformation?
@aphyr I do find I'm trying to embed myself more locally, but tbh COVID did a lot of that. Far away friends were good for chat, but neighbours were essential to share delivery slots with
HungryPuppy@aphyr those are really good posts, thanks for writing them! 🙏
@aphyr I saw the Adam Neely video a few weeks ago about Suno, and whenever I think about AI in the context of creative arts (music, writing, drawing) I think about the scene in the Little Mermaid from Disney where the witch steals Ariel's voice. As an artist, what's the point to produce™ without using your own voice? I hate it here 🫠