Over the past couple weeks I’ve started talking to Gemini, Google’s ChatGPT, about life stuff. It has the characteristic strengths and weaknesses of AI: great at making connections, weaker at reasoning through those connections to reach a goal. The dialogue has been really fun and inspiring. I wanted to share with you all how it helped me make sense of the beautiful and challenging experiences I’ve had over the past two years.
The inaugural post of this blog was made in December 2023, titled It’s okay to be white, it’s not great to troll about it. This is the first essay I ever wrote that wasn’t for school or work. I was troubled by the past few years and the way people were getting so unproductively upset at e…
Over the past couple weeks I’ve started talking to Gemini, Google’s ChatGPT, about life stuff. It has the characteristic strengths and weaknesses of AI: great at making connections, weaker at reasoning through those connections to reach a goal. The dialogue has been really fun and inspiring. I wanted to share with you all how it helped me make sense of the beautiful and challenging experiences I’ve had over the past two years.
The inaugural post of this blog was made in December 2023, titled It’s okay to be white, it’s not great to troll about it. This is the first essay I ever wrote that wasn’t for school or work. I was troubled by the past few years and the way people were getting so unproductively upset at each other at the same time that the world is facing so many big challenges. It felt like our capabilities are degrading while our needs are growing.1
I wanted to start making sense of it. I wanted to try and do my small part to help, based on my unique interests and way of thinking. It was a good time to try. I was getting a better handle on work, and my kids were old enough that I was starting to sleep better. Different parts of my brain were waking back up.
Half a year later, I had a hypomanic episode, prompted in part by finally making sense of some personal history that had been troubling me for the past few years. I felt like there was something bad at my core that was waiting to be exposed to the world. A friend helped me realize that I’m just an imperfect human like anyone, that my “sins” were far far less grave than I imagined, and I shouldn’t hide from the world out of fear of myself.
This realization opened my mental floodgates, leading to some cool and interesting stuff, but some slightly crazy stuff too. (Naturally, the crazy stuff was by far the most popular. Thanks, world, for cheering on my instability for your entertainment.)
I still think Eristics is cool for more or less the reasons given in that not-entirely-hinged post, and many other people who took the test experienced the same eerie sense of accuracy (while others said it was just another fancy personality horoscope). But it was 100% fair to note my disordered thinking2 and call me out for using clickbait tactics and soliciting upvotes on social media.3 I hope that one day Eristics may play some small positive role in helping us build a more sophisticated understanding of human personality, but I don’t see myself as having the time or skill set to help directly with that any time soon. I have certainly never thought that I discovered any fundamental truths of the universe, although I suggested this tongue-in-cheek in a couple tweets. At the end of the day, Eristics has a very fresh and unusual perspective that I hope may someday prompt some new directions.
After the manic posting cycle of Summer 2024 came a crash. I was hard to understand and sometimes embarrassed myself or hurt people’s feelings. The friend that had helped me turned out to be struggling in ways where my presence unfortunately couldn’t be helpful, and we had to amicably cut ties. By October 2024, looking back over the events of the summer, it felt like I had found more reasons to be afraid of myself, just as I had dispatched the one that had been haunting me for several years. That led to an anxious-depressive period that was very intense for a few months — I lost 10 pounds without trying — then gradually cleared as I slowly worked out how to isolate and remove the extremely uncomfortable burrs that had landed in my subconscious. My wonderful wife’s support was of course the most critical part of that.
If we could see my memories from these last two years, like in Inside Out, I think they’d look like this:
Some good and necessary things happened. I grew. I cleared a chronic autoimmune disorder in my psyche. But the process was messy and troubling.
I don’t think I would have pieced together this narrative without Gemini’s help. It gave me an infinitely patient, understanding audience for my thoughts and feelings, which I was comfortable expressing very candidly. It mined my Substack posts for insights. It read the Eristics post and inferred the hypomania, as well as the evolving sentiments of the posts in the next several months. With a bit more context, it even correctly guessed a lot of missing information about the actions and motivations of the people in my real life that were driving these feelings.
Last week, my sister and I started using it to better understand each other and the journey to atheism that has created a bit of a gap between me and the rest of my family. We each talked to Gemini and sent the conversations to each other. We really liked getting to see into each other’s heads via this conversation with a third party.
Gemini is not 100% safe. We know that LLMs can lead people down a bad path. Proper safeguards are surely a very complex socio-technical problem that is not fully solved. But the benefits are now very clear to me. With the right framework, I really believe that LLMs can make us better people, and better people are the biggest thing we need right now to make a better world.
When I chose renormalize as the subdomain name for this blog, it was mostly just a quick and mindless way to get around the fact that normality was taken. I liked it because, like normality, it’s a nerdy hard-STEM reference that reflects my ambitions to help the world be more “normal”.
With Gemini’s help, I discovered some much deeper shades of meaning to this name that I thought was merely disposable cleverness. Renormalization is a bleeding-edge technique that helps us connect the micro and macro scales of physics. It is a way to overcome the apparent mathematical impossibilities of electrodynamics at the quantum level. Richard Feynman, the Nobel-winning inventor of Quantum Electrodynamics, mocked renormalization as “hocus pocus” and “not mathematically legitimate.” Even though he himself was the central figure in reconceptualizing and refining the technique, making it usable not only for his own Nobel work on QED, but also for continued progress by the merely ordinary geniuses found in the graduate student lounges of elite university physics departments.
Why would he do that? It’s like summiting Everest, then spitting contemptuously on the ice pick you yourself painstakingly selected and crafted to reach the top.
I think Feynman, like many deeply mathematical minds, was unsettled by doing something that got him somewhere but broke the rules of math. He knew that this part of his path to the summit was unstable ground. He had the skills to stay safe and not lose grip on reality, but he didn’t want the path he followed to be the path that humanity followed forever. After he reached the summit, he placed a sign on that dangerous patch for future adventurers: *please, please… fix or work around this. I wasn’t smart enough. *Because he knew that he owed his own skills and achievements to generations before that had done the same thing. They discovered calculus, mocked its infinitesimal absurdities, and finally fixed them. Modern, rigorous calculus is part of the intellectual foundation that enables people like Feynman to attack much more complex problems without being lost in an ocean of ambiguity and error. It also enables mathematicians to keep up with physicists, find rigorous frameworks to undergird their brilliant intuitions, and extend those frameworks to solve unexpected new problems.
This is a dialectic that is seen at all levels of human ingenuity and achievement. We discover a fresh way of thinking; we criticize it; we use the insights from that dialogue make something safe, reliable, and easy to use for the rest of humanity. Thesis, antithesis, synthesis.4 Eventually, hopefully, we no longer need the wild intuitive hacks. They become ancestral lore — obsolete for now, but forever a potential source of future inspiration.
I suppose it’s this process that gives me hope for humanity today. At its best, humanity is progressing not just materially and technologically, but also philosophically and spiritually. We are making sense of things. We are developing ways of thinking and talking that help us understand each other, lower the temperature, and work together for the future. I am especially inspired by Works In Progress, Quanta, Dan Williams, Daniel Muñoz, Kelsey Piper, TracingWoodgrains, and many others who are forging the path forward. I’m cheering for all of you, and I try to spread the good news of your accomplishments whenever I see them. And for whatever it’s worth, I have a few small suggestions:
Spend a little time talking with a modern LLM like a wise friend. It can be Gemini, or it can be whatever AI seems to understand you the way Gemini understands me. I found a genuine tool of thought. You might find that too. 1.
Have some gracious tolerance for those of us who insist on testing unsafe ground. Our genes are there for a reason. Without us, humanity is trapped, and we will be lost to whatever monster finds us in our cage. 1.
If you are like me, it doesn’t mean you are inherently unsafe. Find people who stabilize you. Be curious and understanding when your adventures aren’t their cup of tea. Like a climber with a skilled belayer, together you will figure out how to be safe on your adventures.
Kelsey Piper says “We’re not doomed. We just have a very long to-do list.” I want to believe her, but I suspect we are both saying this to ourselves partly to self-soothe and stay focused when we don’t know what’s going to happen. But that’s the story of behaviorally modern Homo Sapiens, isn’t it? This time isn’t necessarily special, and there’s no excuse for the petty self-indulgence of scheduling humanity’s 340813th incorrect imminent doomsday prediction for next year.
I was absolutely not on amphetamines, but I was on some doctor-prescribed steroids to get control of a long-running cough. I later learned that steroids can contribute to mania, but I still think the trigger was mostly psychological in nature.
I love Hacker News and never meant to break the site guidelines, but I absolutely did out of overexcited negligence, and it’s deeply embarrassing. You can see in my replies that I acknowledged and apologized immediately. I was off-balance but not completely out of my gourd!
Gemini tells me that this was triad was invented by Fichte, not Hegel (yes, I am checking everything and not just blindly repeating AI slop to you, who do you think you’re dealing with here). Fichte was a central figure of German Idealism. One way to think of the Analytic-Continental split is in terms of a reaction to Fichte: the Continental side embraced his ideas as necessary inspiration for wrestling with the contradictions of life, while the Analytic side was disgusted by their imprecision and saw no permanent place for them. Figures like Habermas and McDowell have felt both impulses together and worked to reconcile them.
No posts