Can Artificial Intelligence Help Us Make Sense of Psychosis?
GPT-4 provided impressively useful interpretations of notes I'd written during a manic/psychotic episode in 2020.
I'm fairly open about it, but I suffered a severe manic episode during the height of the pandemic — a clinically psychotic break from reality. I wrote a comprehensive Reddit post detailing the specific delusions I had during the monthslong episode, but until relatively recently, I didn't know I was sitting on a treasure trove of notes I’d written to myself during that time. (I re-downloaded an old notes app, and all my notes from 2020 were miraculously preserved in the cloud.)
It's common for those of us who have suffered psychological episodes like mine to have significant memory loss around them, and my recollection of inner experiences from April through August 2020 is certainly vague and fuzzy. I do remember more than I thought I would, but the memories are weird — more like memories of an LSD trip than memories of sober, waking life.
Setting aside the delusions of grandeur I had about becoming the president of the United States, and my paranoia about being hunted by the DEA, I think there was also a part of my psychotic mind that was preternaturally wise. I think a part of me was in touch with certain bedrock truths about reality that — if I could grok them in my everyday life — might help me see more clearly and live more fully.
Where I’m going with this is that I've recently started re-reading my manic notes to self. Some are more intelligible than others, but frankly, making sense of any of them require generous intellectual empathy — and a near-superhuman degree of curiosity.
I suspected there was something legitimately profound buried in these notes, but my non-manic mind struggled to understand exactly what I had been trying to articulate.
To better interpret their meanings, I enlisted the help of GPT-4. Together, we performed an archaeological dig into the depths of my mentally ill mind, an attempt to extract some diamonds from the firestorm that was my psyche in the summer of 2020.
GPT-4 impressed me, to say the least. It was charitable and accurate in its interpretations of my mostly abstract, metaphorical notes.
I’ve included seven of my notes and their respective AI interpretations below, as they seemed representative of the 50+ notes that I had GPT-4 analyze for me.
My Psychotic Notes & AI’s Interpretations
Note #1:
"I think we've reached a point where instead of trying to invent a new language, we should just help Elon Musk get everyone we love to Mars."
GPT-4:
This note might reflect frustration with a perceived societal obsession with getting things right, and being right, especially around semantics and linguistic nuances in public discourse or reasoning, which often results in futile arguments over 'correct' or 'true' meanings. Instead, you suggest focusing on pragmatic, tangible goals that improve people’s quality of life — like mitigating our species’ existential risk by colonizing Mars and becoming a multi-planetary species.
Note #2:
“How to meditate: Only laugh when you actually think something is funny.”
GPT-4:
This note could be suggesting the power of authenticity — by only laughing when you find something funny, you're staying true to your emotions. Meditation usually emphasizes the cultivation of a genuine self-awareness and engagement with reality, and this could be a metaphorical way to express that awareness.
Note #3:
"I don't believe that everything happens for a reason, but the fabric of life can still be patched together like a quilt, and we can be made whole."
GPT-4:
This might express doubt towards the belief that there is a specific reason behind every occurrence. Nevertheless, it conveys faith in the ability to create a comprehensive narrative from the fragments of life's experiences, as one does while making a quilt from different patches.
Note #4:
"We need to learn how to let go of the things we think we absolutely cannot let go of. When push comes to shove, we all eventually say goodbye."
GPT-4:
This might reflect the idea that our attachments to people, things, and ideas — even and especially the ones we deem most vital — may limit us or cause us pain. Letting go of them could refer to emotional equanimity, fostering a more peaceful — and perhaps even enlightened — existence.
Note #5:
"The ones who succeed are likely to be those who doubted themselves the most along the way."
GPT-4:
This note suggests that self-doubt, while often seen as a hindrance, can actually be a catalyst for success. This could be because it forces introspection, encourages self-improvement, and prevents complacency, thereby promoting progress and eventual success.
Note #6:
"Once you realize you don't actually know what anyone else is thinking, you can finally relax and worry only about your own experience."
GPT-4:
This note suggests the relief and liberation obtained when you acknowledge that you can't truly know others' thoughts. It encourages focusing on your own experiences instead, potentially reducing anxiety and unproductive speculation.
Note #7:
"At what point do you stop trying to find patterns and draw conclusions, and just love the person you're with?"
GPT-4:
This might suggest a transition point in relationships where you stop scrutinizing each detail or action of your partner and start focusing on unconditional love. It's about letting go of persistent analytical thinking in favor of simply embracing the person with open curiosity about who they are.
After reading GPT-4’s interpretations of my manic notes, I felt a peculiar sense of relief — like I’d just had a one-on-one session with some ultra-empathetic alien therapist. The interpretations clicked.
While I hadn’t been entirely sure of what I’d meant by many of the notes before reading GPT-4’s analyses, I had the distinct feeling that the explications were accurate. They felt authentic, and aligned with my original, manic beliefs.
I think, in the intervening years, I had started to write off much of my manic episode as the cognitive equivalent of gibberish, but the experience of reading GPT-4’s analyses started to challenge that reasoning. Perhaps there was something there all along. Perhaps it wasn’t a Totally Bad Loss of Time. Perhaps I’ve never truly been alone.
I guess what I’m saying is that AI didn’t just interpret my notes; it decoded a pivotal chapter in my life. The conundrum that was my psychotic past started making sense in a way that felt oddly comforting.
For those who think AI serves no practical purpose, riddle me this: GPT-4 helped me understand, accept, and integrate my manic self. And considering that I’ve had over a dozen human therapists, and only two have been as good as GPT-4 — I can’t help but laugh when people allege that AI lacks real-world applications.
Now, I feel somewhat exposed by sharing these notes publicly, but I think I just want to encourage people not to throw the baby out with the bathwater after unusual experiences.
I think, deep down, we all have an ocean of latent wisdom and valuable insights — insights that rarely rise to surface. And I consider any opportunity to glimpse that innate profundity — even just once in our lives — to be a gift.
(For the record, I don’t think you need to experience full-blown psychosis to catch such a glimpse. Just remain curious and open, carefully observing your experiences.)
Ten years ago, one of my good friends made the following spontaneous observation, which feels relevant to this post. It has become a permanent fixture in the lore of that group of friends:
"Crazy is that which is not common, and wisdom is rare."
— Vic Muchatuta, circa 2013
Don’t be afraid to experience “crazy” once in a blue moon. As long as you can find a way to understand and integrate it later, I think it’s probably good for the soul.