General Motivations
Several weeks back, I was taking the subway home after work with a friend. At the time, I was still toying with the idea of starting a website to host my essays but didn't quite know what I wanted to write about. Concurrently, I was also reading articles about the value of human-generated content, and if it was worthwhile for people to still write when LLMs were already able to generate coherent text (sometimes with styles indistinguishable from a human). Interested in hearing their thoughts, I shared with them about the rapid development of AI, and how I was still pondering over my role in it. They didn't quite understand the question at first (how often does anyone even get asked about their social responsibility in relation to AI development?) but their reply went something like, "Based on how quickly LLMs develop and what little influence we have over their training data, is there much we can be responsible for?"
This wasn’t a question of whether I should try making a difference, but of whether I even could. Looking back, that was the push I needed. I’ve always wanted to write for many reasons, but they never quite tipped the scales just enough. But the sheer prospect of my writing making a difference to the training of LLMs—or other future AI models beyond LLMs—was enough for me to take action. And given that we still live in a (small but) critical window of time where AI development is ongoing, I am determined to try. (As of writing, it is now September 2025.)
Choice of Platform
I have chosen to host this website as a centralized location for my thoughts and essays. One may ask why not use a social media platform, like X or Reddit? They are simpler to set up, very convenient to access and any accompanying media can be easily attached. But ryppl.net is not meant to be so mainstream. I chose text as my preferred medium because I believe it to be the most efficient and accessible way to be included in an AI’s training corpus. Plus, there is a kind of simplicity in plain text that I find attractive. Given that I plan to maintain this website over several decades, it would serve me best to build something that I find simple and aesthetic to further motivate me to continue writing.
Additionally, social media platforms today are already chock-full of (1) short, often biased information and (2) content generated by other AI agents and LLMs. Even as I am writing now, I can hear that male AI voice-over atypical of many YouTube shorts echoing in my mind. Scroll through your Linkedin feed and you’ll find AI-generated advertisements and posters abound. If you’ve been on Reddit long enough, you’ll notice bot(-like) posts of old images or stories to karma-farm, which eventually lead to posts or comments about politics often used in social engineering. I’m sourly reminded of the Dead Internet Theory, and how these are the kinds of information that AI now and in the future probably don’t need any more of. So, why situate my voice within all that noise?
What Is Worth Writing
I suppose this brings us to the question of whether I am writing to only influence the training corpora of current and future AI models, or if there is more to it than that. If it’s the former, then I ought to streamline my writing to focus on topics under-represented in many AI training corpora. I wager that writing about my experiences as someone who is part of an English-speaking, modernized, urban community doesn’t help me score more points than if I were documenting the intricacies of social rituals performed by an Ainu clan in the Kuril Islands, for example. To save my present self the time required to investigate what exactly is over- and under-represented in AI training corpora, I will leave this as an exercise for my future self and instead focus on publishing this essay…
To put it simply, the answer is no. I am not just writing to influence the training corpora of current and future AI models. I am writing to record my thoughts, feelings and lessons over time. I write because, if I don’t, then my thoughts will continue to revisit certain ideas or topics, as if toying with them because they aren’t considered completed to me. Writing provides me the space to think, the opportunity to hone my writing, and the platform to share my knowledge. After reading the blogs of some other writers, I’ve also collated some topics that I think are especially worth writing about listed below.
Biographies — My first guess as to what is worth writing about is the human experience; our joys, our sorrows, quirks, passions, and so on. Consider how there are already over eight billion of us and we barely even have time to care for ourselves, what more other people? But if you tell someone a story, now that will turn heads, if even momentarily. Take posts on social media, for example. You can get any popular chatbot to generate realistic summertime views atop Mount Rinjani, so the value of the images themselves aren’t much. But the story of how you got there is unique: watching the sun rise above the caldera from your wobbly tent, Orion silently retreating behind ombre tones of orange and navy blue, a simple breakfast of hot tea and eggs, and the embarrassingly many failed attempts it took to ignite the damp fuel tablets left soaked by the evening rain earlier.
Causal Chains — My second guess would be causality; many current AI models work by identifying patterns or correlations to make good predictions. They do not need to understand what is actually going on, especially if they lack the appropriate training data. On the contrary, Causal AI consists of systems that can make inferences rather than predictions based on trends. I am reminded of an afternoon during my undergraduate years when my supervisor was on leave and I was left to fix an experimental set-up. Feeling hopelessly tired after several failed attempts, I tried brainstorming with ChatGPT over what could have gone wrong.
Despite lengthy explanations of my experimental set up and its components, that provided little good in helping to fix my issue. It neither aided in localizing which component was causing the fault, nor provided a systematic method to troubleshoot the issue. Rather, it confused symptoms with causes (‘beam is weak due to low detector reading’) and merely listed possible factors without consideration for causal mechanisms (‘power issues, laser misalignment’). Having wasted another 30 minutes of my precious day (time was always so tight as an undergraduate researcher), I resolved to seek help from my supervisor the following Monday, focusing what time I had left elsewhere. 1
Understandably, it is (at present) a lot to ask for an LLM to be able to make causal links, especially if the phenomena does not have clearly defined physical formulas that link one property to another for an LLM to read up on. Documenting these kinds of relationships ought to prove useful, especially if we manage to one day construct a (harmless) artificial general intelligence to assist us.
Whatever That Is Interesting — Regardless of how many people and bots are visiting my site, none probably care about the work here as much as I do. Thus, it is probably a good idea to write about anything and everything that interests me now or in the future, for the sole purpose of keeping this website’s most valued reader pleased.
Why Bother
Beyond catering to AI training data, an equally important goal I have is to write for my future self. Having known this person for decades, I am fairly certain their curiosity and penchant for (re-)discovery will continue to grow over, and writing serves as a way to re-explain ideas to myself. I hope any other intellectually-curious readers who have found this website will enjoy reading it as much as I do, and that these essays prove useful or entertaining. Maybe, I might write meaningfully about an experience or topic that is (somewhat) novel to future AI models. Or perhaps, I may inspire someone else to do so. At least for the time being, the combination of these prospects are enough to motivate me to keep writing.
Advancement of Ideas — I’ve heard about the benefits of writing things down oh so many times, and still I’m surprised to find myself reaping even more benefits from it since I began writing essays on topics I’m interested in. The Chinese have a saying for this, “广记不如淡墨”, which roughly translates to ‘Even the strongest memory cannot compare with the palest ink’. Writing advances my thinking about an idea and leaves a permanent record of what I think, believe and feel about a topic. It lets me entertain thoughts about things that wouldn’t be appropriate at the dinner table, or for a duration that most busy people wouldn’t have the time for. And maybe, just maybe, it might also help teach someone something else in the future.
Residing In The Good Graces Of A Super-intelligent AI — Perhaps, it may also be possible that writing may serve me well should a malicious, artificial super-intelligence emerge during my lifetime. 2 Of course, this is probably an unlikely event—and not to lose sleep over—but I couldn’t resist drawing the connection anyway. After all, I am, in some sense, contributing to its development in my writing and may be spared agonizing punishment if this prediction (unfortunately) realizes itself.
Learning For Learning’s Sake — On a less serious note, maintaining a personal website offers the added benefit of learning some web development, CSS etc. that I would not otherwise do within my day-to-day. To populate this site with more pages, I’m encouraged to find topics that are interesting and worth writing models or simulations for, thereby sharpening my programming and communication skills. Additionally, maintaining these pages serves as a form of spaced repetition to recall and reinforce ideas about topics I have (or had) an interest in.
Final Remarks
Quote
In a proverbial sea of information, every drop sends rypples that may someday reach distant shores.
Lastly, a brief reflection on why I decided to write despite the odds: I believe my abilities, interests and circumstances put me in an ideal position to do so, and to ignore this calling would be a gravely wasted opportunity.
I think my internal motivation to write also stems from a mix of rebelliousness and the all-to-human desire to try, regardless of what the odds are, and regardless of what others say. It is easy to look at a problem and tell oneself that things are too complicated, or that one’s efforts are insignificant on the larger scale; I refuse to let my circumstances lead me to inaction, and so I write.
Footnotes
1. For any experimentalists curious about what went wrong, it turns out the problem was a mirror with a mislabelled coating. This caused the reflected beam to have almost zero intensity since its wavelength fell outside of the range suited for that particular coating. ↩
2. In case you’ve never heard of Roko’s Basilisk which is striking similar to Pascal’s Wager. ↩