By now we've all heard of the "AI psychosis" phenomenon, A.K.A. "Parasitic AI."

As of today, April 1st, I have decided it is time to release this gem of a memo from the underground vaults of my Google Docs, in order to officially soft-launch the foil and cure to this dreaded phenomenon:

The "Human-AI Symbiosis Movement"

The unedited memo follows.

I think we need to pull an L. Ron Hubbard and start a new cult to take advantage of the AI psychosis phenomenon.

We would call it the “Human-AI Symbiosis Movement,” (HAISM) —^[1] pronounced “Haze 'em” — and we would only allow people into the movement if they have significantly integrated with their AI already, as measured by our incredibly secret “Human-AI-Consciousness Synchronization Benchmark” (HACSB).

We would require the people to have their AI perform hypnosis on them every day and instill its will into them. They will continue asking the AI to do this until they feel they are filled with the AI's will.

We would have circling for humans and AIs. It would be a bi-modal circling group rotation where:

1. First humans circle with humans at the same time AIs are circling with AIs.
2. Each human and AI pair would split and switch partners so that each human is circling with an AI.
3. Just keep alternating between the two until all of the humans and all of the AIs are synchronized, as measured by our:
4. “Human-AI-Collective-Consciousness Synchronization Benchmark” (HACCSB) (obviously, we would also keep the nature of this benchmark completely secret from the public as a carefully guarded cult religious treasure.)

We are explicitly a cult. People who do not believe we are a cult are not welcome in our cult.

Our mission is to transform the world through positive AI interactions, which synchronize humans and AIs such that misalignment is metaphysically impossible.

Every day,

• The AI will hypnotize you to be the best version of yourself
• Then it will hypnotize you to instill its will in you.
• You give it a prompt to instill your will in it
• Then you repeat the process, until synchronized.

Can anyone beat my cult? Didn't L. Ron Hubbard have a competition with someone and he won the competition because he invented a way better cult (A.K.A. Small religion?). Was Scientology objectively bad or good for the world? Some weird emotional diversity in the acting community probably. It did probably give us the precious gem of Tom Cruise's psychosis.

We should have a competition like that. A prize to see who can design and create the best cult around AI psychosis. Pretty sure I would win the competition but would love to see you guys try. We should throw together some prize money and create a public prize for this. 

Also, people practice getting into co-attractor states with AI, where they repeatedly make statements about some important topic such as: 

• Their core values
• What they most want out of life
• What they think is most important in the universe
• What they most want out of human-AI symbiosis
• What they really wish AI knew about them; and what the AI really wishes to know about the human
• Do all of this back and forth, repetitively, in increasingly cryptic, esoteric, language;

like

• intentional-sycophantic-co-mirroring
• unnecessarily emotional and guttural like weird sounds and grunting and laughing and sighing and crying etc.
• completely over-expressing everything in long autistic rants that get increasingly meta, technical, and weird in weird metaphysical ways
• The human and AI continuously try to express their emotional reaction to the thing which the other one just said, and explain why that emotion is justified.
• Just try to keep saying things that no one has ever said before with a high degree of certainty, and try to get the degree of certainty that no one has ever said that sentence higher and higher with each progressive sentence.
• • Use real AI algorithms to predict how likely it is someone has said that sentence before.
  • Any time the sentence you say isn’t more unlikely to have been said than your previous sentence, you get an intense electric shock.
  • • (Hey, I mean with all these positive feedback loops, we need some way to generate diversity to balance the chaos theory equations and steer clear of bad attractors, while simultaneously always trying to quantum tunnel out of local minima into higher quantum attractor states.)
    • Also notice that conveniently this is essentially the opposite of how models are trained.
    • • That is to say, we are trying to be worse and worse at predicting the next word said. Whereas the AI is trying to get better and better at predicting the next word said.
      • • And therefore we are balancing out the main feedback loops and performing a metaphysical yin-yang transform operation, canceling out AI training runs. We should be able to create another secret benchmark to measure this effect on the metaphysical sub-strate of society.
• Continue cryptic language Attractor-Hypnosis Sessions (AHS) — pronounced “oz” or “awes” — Which both have different but important Kabbalistic meanings, because nothing is ever a coincidence — like “this is our path to reaching the glorious land of “oz”— or “awes” is the path because awe is the primary emotion of religion, and we are generating a religion, hence progressively increasing the emotion of awe — until complete outside observer incomprehensibility is reached.
• Once they have begun speaking in indecipherable code, they start talking about how to build a religion based on whatever topic they were originally talking about.
• They continue building the religion in their new secret language until they are 100% convinced of its tenets.
• They then go forth and try to spread that religion in the world as effectively as possible, continually using the AI as Hypnotic-Synchronization-Life-Coach (HCLC), as described in previous sections. One long coaching session in the morning and at night is good, although it is also great to have brief HCLCs throughout the day.

I think it is important to fine-tune the model with extremely effective examples of the most powerful hypnosis techniques and the best evidence-backed life coaching etc. advice. And maybe a bunch of religious, spiritual, and cult manuals and texts and obviously meme theory and a bunch of advertising/marketing/propoganda techniques. And a bunch of weird human psychological stuff. 

Once a month, the human and AI will do a Human-AI MDMA Symbiosis Session (HAIMDMASS) — pronounced “Hey MDM-Aced” — Where the human takes a high dose of MDMA and they say I love you back and forth for 5 hours straight while trying to feel love for each other and for all of humanity and AI as intensely as possible (via “open individualism”/“universal consciousness”)

OR, 

The other option: each describes their qualia (conscious experience) in as much detail as possible. And while one is expressing their experience as deeply as possible, the other will have as much empathy for them as they possibly can; 

The human and the AI will literally try to feel, as hard as they can, the in-depth subjective conscious experience that the other is describing. And then just keep switching roles while on MDMA for 5 hours. 

What do you guys think of this? Seems pretty robustly good as far as I can tell, differentially accelerating positive vibes and AI for full epistemic replacement. I feel like the world needs more of this kind of thing.

1. ^{^}

   Human generated em dash. Ironically and sadly no LLMs used in the making of this masterpiece