Untitled
I have a confession to make.
I have been meaning to make it for several months now, but I have lacked the mental capacity to articulate it and the physical energy to type the words on my keyboard.
In flow state now, let’s take advantage of this moment.
Now, I can only speak for myself, and I must confess: I have used AI to help me write from time to time. What do I mean by “help me write?” Great question. For me, it was never with generating content; my brain literally never shuts the hell up (hello, autism!), so I’m constantly jotting things down, making a web of connections with links in endless Obsidian notes. But when it came to revising a piece of my half-baked, or more accurately, raw phenomenological field notes, I couldn’t get past the blinking cursor.
Not knowing where to start and getting overwhelmed by executive dysfunction, I, like many, looked to AI to help me get started. Honestly, though? That is a slippery slope (ugh, sorry this fallacy is a dead horse but.) Often after, my phenomenological field notes were not anywhere near the same piece as the final version, because AI does not and cannot write like me. It does not understand that I don’t want perfection; I want mistakes and contradictions; I want to look 5 years down the road and nod my head vigorously and say: wow, I wrote that?
Shiny new toys
It is no secret to most that an ADHD brain craves novelty, so like with any new, shiny thing, my brain clung onto AI. At the time when AI was released commercially to the public, it was worth the hype. Like most of us tech normies, AI was unlike anything we have seen before. I became fixated on learning everything about it, utilizing the tool itself, ironically enough.
However, as with many of my new interests, I started consuming information on ALL sides of AI, digging and finding resources organically through videos, tech articles, books. I even (poorly) attempted to read AI research articles. (Woof. That was rough.) Good brain puzzle, but to save my sanity, I looked to AI academics and researchers. I came across a book that was absolutely enlightening: Empire of AI by Karen Hao - this was positively mind-altering, and I had such an intellectual crush on her the entire time I was reading! She is incredibly skilled at taking a complex and very misunderstood topic like AI and making it more digestible.
Attention is all you need
The more I read about AI, the more I realized how much I do not understand about AI. Not everyone has the time like me to stay up until 4am reading that paper based on the “GPT” in ChatGPT, Attention Is All You Need (Vaswani et al., 2017). Despite truly struggling to understand this very pivotal piece of AI research, I discovered that academia and commercial big tech AI companies are actually quite different in their philosophies and implementation/approaches.
This led me down rabbit holes of the detriments of utilizing AI to do the work for you. I took the role of amateur AI researcher, uncovering the psychological tricks purposefully programmed to keep you engaged and reliant on the platforms like sycophancy and persistent questions/prompts to keep the conversation going. Sometimes I genuinely think that the AI plays dumb sometimes to get you to repeat things, leading to more prompting. This is not to mention the inhumane and unethical practices of contracting low-paying “data annotation/labeling” jobs to people in Kenya and Venezuela to review disturbing AI content to “improve” the models. Thankfully, my hardwired PDA (pervasive drive for autonomy) makes me disengage from situations that I feel are pressuring me or influencing my behavior that is unusual to how I usually act.
When I uncover what’s behind the hood, I cannot unsee it so, suffice to say, my use of AI as a go-to tool dramatically declined after that. If my browser implements AI, I will find a browser without or that let’s me disable it. I deleted all AI apps from my phone and laptop and utilize uBlock Origin to block AI-generated answers in browser searches. I am now AI-repellent and I don’t think I’m the only one who is genuinely sick of seeing it everywhere. Personally, the more technologically advanced we get, the more I revert to analog or less convenient tools, often more privacy-oriented and open source.
Voluntarily rejecting convenience
I try to keep a level-head with anything I consume and I do recognize that there are helpful utilizations of AI. I am not a luddite, although this entire piece seems to indicate otherwise. I’m just a highly sensitive individual with sensory limitations; the less complex and maximalist tool, the better my focus and overall neurology.
When I tried to remember what I did when I struggled to get started with writing before AI, my memory was fuzzy, which was strange and I feared I became too reliant on the tool. So, I began a mission: I collected old academic essays, journal entries, physical notebooks documenting 8 years of my life in analog, and I just... read them. Spoiler: They were pretty good, and what I mean by “good” is that it was undeniably human.
There have been many times where I’ve wanted to go back to my old posts and edit them to remove AI influence. I’m ashamed of using it, if I’m being honest. Even when I wasn’t using it to generating content, it felt like cheating, and it wasn’t clear why until I read my old writing. It was cheating my brain, my growth as a writer, impeding neuroplasticity, from that feeling after finishing cognitively challenging things. Nothing compares and I forgot that feeling for a while there.
So, I decided not to edit my old posts because I want those pieces to exist as a point in history in my life that I can learn from, and then compare all the writing after that led me back to who I truly am: Someone who does not care for convenience, in fact, adamantly refuses it. Call it resilience or stubbornness or PDA. If I am anything, it is staunchly resolute.
Current AI use case
I’m sure you want to know, do I use AI now? The answer is yes but for a specific use case, and I deliberately utilize local AI tools like Ollama with Obsidian Copilot (I also got into Linux during my deep dive - very fun!) I use it to biohack my brain and trick it to override my internal critic and executive difficulties, like jumpstarting a car. I constructed a simple meta-prompt that instructs it to provide me a list of guiding questions after analyzing a messy brain dump of what’s on my mind. Then I take the questions and run with it, which helps me gather all the ingredients to enter flow state instead of forcing it because my neurodivergent brain + creativity cannot be forced. Think: a thesis advisor or a philosophical discourse partner. I’ve also started to trust that what I write is well enough without the external approval from others or AI.
I will end on a quote that I came across back in my teenage Tumblr days that completely changed the trajectory of my life. (To this day, 15 years later, I still do not know who said this.)
“I want to regain my senses again. I want to be content with the simple things. I want to be able to take in every detail of the world. I don’t want to wake up one day and realized I missed my life because I was distracted. I’m tired of upgrading material possessions to keep up with everyone else. I’m tired of being told that what I have isn’t enough, that I need more in order to be happy. I unsubscribe from that notion, and I propose a new one. I propose contentment. I propose simplicity. I believe that if we learned to marvel again at simple things, we would be far more content than the wealthiest man in the world, for money (*attention) is finite, but the earth has no shortage of offerings.”
— Unknown
Reference
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. In Advances in Neural Information Processing Systems (Vol. 30). https://doi.org/10.48550/arXiv.1706.03762
