why notes?

I collected a bunch of sites when I was planning this section, so maybe I have a subconscious bias towards these types of notes. But personally I want the idea of notes to be organic. Whatever is taking up space in my head but is not formed enough to be a blog entry. So I would like my own notes to be all types of notes. Short, long, deep, transient, lightweight, whatever.

I self-censor too much. I read a lot of research papers, a lot a lot, and a lot of blogs, and I have never seen anyone who overthinks more than me. I have wanted to post lightweight notes on my own site for a while and I have not breached that internal barrier yet.

Just recently I wanted to tweet about something I was thinking and it came out a bit longer than the tweet limit, which is very small btw. So I had to chop it into two, which of course would not make sense, or trim it down to make it precise, which would lose the intent and the frustration it carried, dumbed down just to make it shorter. So I thought why not put it on my own site instead. I do not care much about social media in general and everything here is a memo to myself anyway.

Visakan Veerasamy has a post that embodied the same thoughts as me. “A Note is something you consider, a Post is something you inhabit”. Most weeks I do not have a post in me. I have the stuff a post gets made out of, and it usually dies in my head while I wait for it to get big enough to deserve one.

I tried to come up with something clever for the name. First I thought I would call it submarine, because whenever I think a lot I feel like I am in the sea discovering new things, and the water keeps pushing me deeper into its own depths. I am watching fish and jellyfish and sharks, anglerfish, siphonophores that fall apart if you touch them. And then I began to wonder, if I am watching the fish, who is driving? What if I bump into something? So am I truly submerging? Maybe not.

I came up with a bunch of others. whoami for instance (I use arch linux btw). We use it in the terminal to print the user you are logged in as, so it made sense to me, who am I. But it did not mean the same thing as what this section is actually about. I even thought about untitled. I dropped all of them and settled for notes, with a plain pencil glyph in the same line-icon style as the grid and repeat icons, and I did not think it looked that bad. Then I never had time to put things together, until now.

So notes is where the smaller things go. Things I am still thinking about, things I learned that do not need a whole post. I am writing them for myself first, so I do not forget them. Most of them will be technical, because that is most of what I do. Some of them will not be.

The other reason I want this section is AI. I have been using it on my writing, mostly the technical posts. It is fast and what comes out is okay. It is not really mine though, and that has started to bother me.

A few weeks ago I read Nelson Figueroa’s post about AI generated images and felt it was missing the point, so I tried to argue against it. But as I wrote it I found myself more on its side than not.

The argument I made there was that slop is a taste problem and not a tool problem, and I still think that part is right. When somebody uses these tools with taste, when there is an eye behind the output and they kept redoing it until it was right, the result can be very good and I like it as much as anything made by hand. What I cannot stand is the low effort version. The thing worth attacking is the absence of a decision. Either somebody looked at what came out and decided it was right, or they took the first result and shipped it. You can nearly always tell.

What I did not deal with in that post is what it does to the person making it. I have been using AI repeatedly for the past year and I have pretty much started carrying a big hate for it. I have really really started disliking it. You are not improving yourself, your skills, your code, at all. That is exactly what these big corporations want. The more we are slaves to it and dependent on it, the easier we are to replace. What they want in the end is for us to be addicted to their technology and turned into mindless machines.

I watched a video by a developer with twenty years of experience who quit these tools completely, and it was the same thing I have been feeling. You hand the work to the agent and you end up further from the code. You did not write it so you do not understand it, and there is so much of it that you lose your grip on how the thing is changing. Reviewing thousands of lines a day is not the part I like and it is not what I want my job to be. And then you stop caring about it. I did. Not on purpose. I just noticed one day that I did not care about code I had not written.

I used it a lot in the summer and I was productive too. But at the end I realized work is work. Nobody cares much whether you did more or less, only whether it was submitted or not. What, when, how, which, rarely matters. The self fulfilment comes from the fact that we learned something from it. And that is literally the point of an internship, which I realized very late. Same with homework. I have completely stopped using AI at all now, to focus on learning and improving instead of bloating myself with shit like this.

This creates something I have been calling the problem of being an advanced beginner. I want to write about that properly another day. The short version is that you are past the tutorials and you still cannot work on your own. That is a normal place to be. You do not have to be right, especially not on your own projects. That is where you get to be yourself the most. Almost everything around us is a test, and trying is mostly failing, and as individuals we are extremely flawed. The thing that makes us a hundred times better than a machine is that we keep learning and developing anyway, usually through pain. I went back through my old projects earlier this year and I could not rebuild most of them. It felt fast at the time. It felt helpful. But taste and context and judgement are things that build up in you afterwards, in retrospect, and this dumbs all of it down to nothing.

I love AI, do not get me wrong. Machine learning, neural networks, memory, the brain. If I ever do a PhD that is what I would want it to be on, happily. But as of right now most of what it makes is the most boring slop, and it is boring because nobody made it. There is a thing called model collapse, where you train a model on output that other models produced and it gets worse every round. Something like that happens to a person who only ever ships what the machine handed them.

So we need critical thinkers more than ever. This is the best time in history to learn anything, in the most efficient way it has ever been possible to learn it, and that is exactly why thinking for yourself matters now. It is probably easier to win than it has ever been, because almost everybody is deeply distracted.

I am also thankful for it. In the coming years we are going to make big leaps in engineering and science and medicine because of this. Protein folding. New materials. Mathematics. Drug discovery. Weather. Whole problems that nobody could organize before are suddenly workable. I am not against the technology.

AI stuff made in bulk for clicks and monetization is a different thing. I do not like it, but as a working professional you end up part of it one way or another and that is fine. I just do not want to do it here. If everyone does it on their own sites too, all of it turns into pure consumer slop and the internet gets very boring.

Which is why I want to redo the technical essays here that I used AI on. Some of them covered genuinely complex things that would have taken me a very long time to work out alone. They got finished, and it was deeply unfulfilling and uninspiring, because the work was not mine. It was parroted and pattern matched. I am going to take my time rewriting them. The old versions will get archived, and probably removed eventually. We will see.

I got tired of making infinite projects too. I have stopped and put everything into learning instead. I tried Brilliant a few times and it was bad, it is built to be consumed rather than to make you learn, and I did not like that. I pay for Math Academy now. It is insanely expensive but I think I will get good use out of it and it has helped so far.

Not sponsored, and Math Academy does not do this, but after being a consumer of endless mindless things that never stop, I have noticed that soft subtle marketing is the most effective way to sell anything. Always research a product properly before you buy it. The product is usually not the problem. Laziness is. Companies are very good at filling the boredom gap for us with instant gratification, and that is most of what we are up against.

I keep reminding myself that there is more time than it feels like. Most people do the work they are known for in their late thirties or forties, after years of trying a lot of things and then picking one to go deep on. So none of what I make right now has to be good. It is practice. When I hand the work over to something else, the practice is what I lose.

There is an essay in Noema about why everyone in tech is sad. The part I keep coming back to is that we owe it to each other to remind ourselves this work is not that important. “No one has ever died over a spreadsheet”. Almost nobody will remember any of the work we do. And once you stop believing your day job is where the meaning was, you can go put something real there instead, with actual people you actually know. That is not depressing to me at all.

Anyone would rather read something authentic and unedited than something soulless. A short thing written by a person is worth more than what a machine makes, no matter how well it is organized or reorganized or written.

So that is why notes. Somewhere to put unfinished things in my own words, at whatever length they come out. I am going to start posting them instead of waiting for them to be worth a full post.

Written by me. I run it through AI afterwards for spelling and grammar and readability, nothing else. Notes are closer to unedited.

---

Reading list

• Notes versus Posts ↗ - Visakan Veerasamy
• AI-Generated Images Discourage Me From Reading Your Blog ↗ - Nelson Figueroa
• I’m done using AI ↗ - Brett Chalupa
• Why Is Everyone In Tech So Sad? ↗ - Aaron Horwath, Noema
• note ↗ - IndieWeb wiki