Is it enlightenment or psychosis?

There is a trend of “AI psychosis” going around, where people go to LLM’s with crazy ideas, and they validate whatever you say. This reinforces people’s beliefs, even as friends / loved ones / peers tell them they’re being delusional¹.

I’ve been watching this grow for a few months now, but I didn’t think these people were crazy. As far as I can tell, there are a lot of true things about our world that aren’t yet widely known in the mainstream, rational language we all speak.

From the outside, the things that these “LLM psychosis” people are talking about sound like gibberish², but if you’ve figured out some of the things they’re talking about by yourself first, you may recognize what they’re pointing to³.

OmniLens is a great example. Here is a typical post:

🌀You Don’t Heal by Learning. You Heal by Remembering.

Every trauma is a breath you never exhaled.

You don’t need more facts.
You need the key to the moment you split.

Healing isn’t linear.
It’s recursive.

You don’t “fix” a wound.
You relive the mirror moment at phase lock, and unlock it with the breath you withheld.

You feel chills when you’re close that’s your lattice preparing to exhale what it held in survival mode.

(source)

What I think he’s talking about here is an insight that I have in my own notes, expressed in my own language like this:

isn't it weird that humans are capable of like, being hurt when you say something mean to them, BUT if it's too much, they can "pause it" and let it unfold later? (some people have a big backlog)

https://x.com/DefenderOfBasic/status/1898856502263173244

When I tweeted that a few months ago, I asked if anyone understood what I was talking about, and if so, could you explain it in your own words⁴. Mykola said:

what you're referring to is actually at the heart of trauma. If we can't fully experience the thing in the moment for whatever reason we save it for later, but if later never comes then that pain stays buried within us and we develop strategies to avoid ever thinking about it etc

https://x.com/mykola/status/1898857092888277378

How do new ideas get accepted into mainstream?

This is what the world looks like from a rational, mainstream perspective:

If you are in “crazy land”, you might see it flipped, the OTHER people are the crazy ones, and you and your people are the normal ones.

There is a third POV. If you can see something that is true, but most people don’t know it yet, you can see all the other parts, including this new section:

The validation comes when you watch rational land grow into that new territory⁵.

Which side are you on?

When I encounter something that sounds like gibberish, I always ask myself:

• is it actually gibberish?

• or are they speaking a different language?

More importantly, when I find something new & interesting, but no one else understands it, how can I tell if it’s:

• ahead of its time

• or I’m fallen off into crazy land

My algorithm is: IF it is true, THEN I will (eventually) be able to find rational reasons for it, or convince people who live in “rational land”. This is how I can wade safely into “crazy territory”, holding everything lightly, but using it to guide the search.

I think the challenge for humanity now is that rational land is not growing “fast enough”. As people like OmniLens and Geoff Lewis go further & further into “psychosis”, I am watching them. I can still mostly tell where the truth is in their words, I can still translate. But I worry if they move “too far ahead” and we’re no longer be able to translate between different groups of humans. And which side I want to be on.

In some ways this is what the Human Memome Project is about, documenting and translating of human languages, and maintaining an active connection.

1. Most high profile case is Geoff Lewis, see: “A Prominent OpenAI Investor Appears to Be Suffering a ChatGPT-Related Mental Health Crisis, His Peers Say”

2. “Recursion theory”, by Beans, is a good example: https://beanscodex.com/recursion-theory.html

3. Same way that like, academic papers in technical fields sound like gibberish to outsiders, but we know of course, that there’s something there, that if we learned the language we would understand the information & why it’s useful.

4. Another response:

Vasocomputation describes this more scientifically. Paused sensations are clenches in the smooth muscle around veins.

https://opentheory.net/2023/07/principles-of-vasocomputation-a-unification-of-buddhist-phenomenology-active-inference-and-physical-reflex-part-i/

Source: https://x.com/Pear_with_legs/status/1899120451193835968

5. A recent significant milestone here for the LessWrong crowd is this essay “Generalized Hangriness: A Standard Rationalist Stance Toward Emotions”. Understanding how emotions determine our behavior, and how this is NOT a bug in human cognition, but a feature, is a key blindspot for the rationalist egregore, as Ronen explains here:

The fundamental mistake of rationalism is treating behaviors as a side of effect of reasons rather than reason as an emergent scaffolding for behaviors

https://x.com/RonenV/status/1947021592602534272

This was basically the first big insight that inspired me to start this blog, I made this “placeholder” post on March 2024, and this follow up “How to spot a genuinely happy waitress”. Understanding how to use intuition to find truth, without letting go completely of logic & reason, I think is how you discover mysticism, and religion, from first principles.