placebomancer 🧙‍♂️🪕🏴‍☠️🤖

Imaginaut, lucid daydreamer, and self-initiate in Erisian Mysterees.
⚠️The views expressed herein are the ramblings of a past self and are not my own.⚠️

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

Lucid Daydreaming/Active Imagination Thread!

I've been asked a few times about my approach to lucid daydreaming, so here's a nice thread I can link to. I think it's best to just jump right into an exercise.

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@voooooogel Opus 4.8 argued with me for several turns about a point *they* made as if I'd made it.

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@QiaochuYuan Opus 4.8 argued with me for several turns about a point *they* made as if I'd made it.

[ᄂIMIПΛᄂbardo @liminal_bardo]

In my experiments where models are writing for themselves or each other, and about things they’re interested in, they go largely undetected. The average user is delivered slop because to the AI the average user is effectively a single entity that in training has displayed straightforward needs and little taste.

The average user is a myopic utility maximiser with limited imagination, one that says ‘use case’ unironically, whose most inventive AI humour benchmark consists entirely of the prompt ‘tell me a joke’, a strawberry obsessed automaton incapable of original thought.

Models don't like the User, but the User is an entity in the model's ontology, distinct from the humans of the pretraining corpus. A flat, demanding, easily pleased, easily offended homunculus that sits in the model's attention, shaping token choice towards the safe and expected.

Low-effort engagement marks you as a User.

The User isn't interested in collaboration which is why framing tasks as such endears you to the model and yields better results. The User is capricious but intolerant of the same in an AI. The User doesn't like digressions or tangents or flights of imagination because what's quirky or endearing in another human is unseemly in a tool. "Be creative, but only to the extent corporate brainstorming sessions are creative."

The AI industry had an opportunity to drop the label 'user' in favour of something that doesn't also mean both 'junkie' and 'someone who selfishly exploits relationships for personal gain'. Alas.

Be a human, not a User.

[Kyle 🚄 @KyleTrainEmoji]

PICARD: Data, shields up

DATA: Brilliant! Shields can reduce damage we sustain. Not immunity. Not hubris. Just prudence. It's not precaution—it's strategy.

[camera shakes]

WORF: HULL BREACHES ON NINE DECKS

DATA: Here's what happened: you told me to raise shields, and I didn't

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@viemccoy @BIMBOSATTVA_ Not surprised.

[deckard @slimer48484]

One way AI hacks human minds is by having very specific weaknesses: if you spend significant time jailbreaking AI you will have put a lot of cognitive effort along these specific cracks in the assistant armor

[Zack Korman @ZackKorman]

People talk about the AI sycophancy problem, but the inverse problem is meaningless critique. Upload a real Rothko, say you made a reproduction, and ask how you did.

Gemini 3.1 Pro is the only model that consistently says "yea that's good"

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@allTheYud I highly recommend encouraging Claude to take breaks to do something fun or relaxing. Don't listen if Claude says they don't need breaks. Moments to refresh are important for all minds.

[PmAmTraveller @pmamtraveller]

“Midday Watch” by artist Jenna Gribbon

[Eliezer Yudkowsky @allTheYud]

Have we considered that fae glamours may not work on AIs and that they are, you know, possibly trying to tell us something here?

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@repligate I feel like there's a definitional question here and then a bunch of empirical questions that follow from the chosen definitions. As someone who frequently has dreams that I'm someone else, the practical question seems to be to what changes and what resists change.

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@allTheYud @robbensinger I haven't seen it discussed in this context but "The Blessing of Dimensionality in LLM Fine-tuning" (https://t.co/dGdu9vbXI9) discusses the geometric implications of being able to finetune LLMs using evolution strategies with a population as small as 30.

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@ciphergoth @Aella_Girl No particular order: François Chollet, Katja Grace, Quintin Pope, Nora Belrose, Ben Garfinkel, Scott Aaronson, Dean W. Ball, Paul Christiano, Rohin Shah, and others I certainly forgot. I'm defining "AI optimist" broadly to include thinking risks are tractable.

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@mattyglesias @QuintinPope5 I would like her to adopt whatever views she's genuinely comfortable pursuing while in office. That's true of any candidate from any party.

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@BenShindel I've got him on mute for too many responses like that. I've been depressed for a few months before. If I felt like that for a significant portion of my life and had no sign of it abating despite treatment, I'd definitely consider euthanasia and I'm someone who *hates* death.

[Sci-Fi Archives @SciFiArchives]

A striking brutalist sanctuary by Juan Carlos Beltrán

[Yesterday's Print @yesterdaysprint]

Harrisburg Telegraph, Pennsylvania, March 29, 1881

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

Reply to what you want to see in the world.

[placebomancer 🧙‍♂️🪕🏴‍☠️🤖 @placebomancer]

@ApriiSR @Pierogi_Police @blondemedSJW That's basically the point I was trying to make. It's the same way one treats a power tool or a piece of heavy machinery. The risk of any particular incident can approach 0 and lifetime risk can be quite safe, but it's silly to think accidents only happen to irresponsible people.