I’m feeling kinda funky so I’m gonna try to keep the idea I’m trying to express sorta simple. (Edit: This is incorrect, I ended up spending 4 hours blogging because apparently two cups of coffee after waking up is too much for me to handle. Sorry.)
This post is written and shared primarily for you, friends / acquaintances / strangers who found my blog via google. But also _partly_ as a reminder to myself.
Sometimes we doodle aimlessly or make funny sounds that feel good in the mouth, or just noodle around on our favorite instrument enjoying the sound. That is MAKING for you, the enjoyment is in the making.
Sometimes we have a specific idea we’re trying to record or express or that we want to witness because the version in our head is kinda hard to pin down into a concrete form that won’t vanish. When we make things to capture that, or make things to try to _find_ that satisfying result, that is making for YOU. The method of getting it made is not so much as important (possibly, depends on the person) as the matter of getting to the right outcome.
And in those particular cases, the act of sharing the thing that was made is usually a separate decision come up with after-the-fact. The purpose of making the thing was already fulfilled, so sharing it with others is either a byproduct of yet another goal, or a habitual activity, or a social connection task, or any number of things. But it itself is an act separate from creating (usually). Picking which things to share and why is a matter of Curation, which is kind of a subtle art of its own. Choosing and arranging discrete choices is art! Even if it doesn’t really feel like it sometimes.
The decision to make things during a live stream or in a shared workplace is kinda merging the act of sharing and creating together, but in that context the constant barely-filtered sharing is the dominant activity that colors any Making that takes place. The act of curation is limited in that context and has to take place as part of the process of creating the thing at all.
And then there’s making things that are intended primarily for a friend, or a client, or to be shared with strangers. The entire value assessment shifts into a zone that kinda depends on feedback of those social connections to others, with a little bit of the personal curation coming through as a matter of what feels good or reasonable to one’s self, and as a matter of considering social consequences of the things being made. When we are Making for Friends or Making for Others, we are almost always using a very different underlying ruleset than we do when we are Making things for Ourselves. Because a lot of things that we can do to get practical information for next steps and then discard once we reach that next step take on a different meaning and feeling when it is instead something that everyone sees and forms opinions about, intentional or not. I might spend a week going through two or three pages worth of funky incomplete sketchbook doodles before one of the ideas latches on in a way that I genuinely feel I want to explore and flesh out. And then even with that one thing I might drop it after a day or two, or I might forget it exists overnight, or I might get really excited for a month and start doodling it multiple times across friendly drawpiles or request streams and the like.
(I never originally intended to share this. I am only sharing this because the sketchbook was right next to me and it’s one of those cases where I just kinda doodle things aimlessly and don’t feel particularly strong about it. In this case the doodles are fine I think, they’re just lacking zest I guess. of my “burn energy by making stuff” doodles, this is definitely on the tamer and more coherent side.)
Anyway, what I’m finding is that while I do kinda like sharing things with friends, especially when they’re made for friends, I cannot thrive on that energy. I *have* to be making for _myself_ if I want to keep making things. But directly contradictory to that is that I do not like the process of making things. If I could make the ideas in my head a reality without having to do it myself, I absolutely would, and I would enjoy the outcomes immensely, and it’d be a losing die-roll for most of those ideas whether I’d feel like I should share them at all or just keep them to myself.
And I feel like that that’s the pandora’s box I’ve walked myself into by becoming more familiar with modern machine-learning technology.
(!!Big Ol Warning here because I’m about to pivot and start rambling about Machine Learning!! and how I used to want to go into that field about 15 years ago and how now Big Tech/Tech Bros and working artists dealing with the newest Capitalist flavored Existential Threat keep tossing out knee-jerk takes that make it impossible for me to share my sincere feelings in any kind of mixed setting unless I want to get attacked by unintentionally misinformed takes. I won’t be offended if you stop reading here, I promise. Okay well I might be _slightly_ annoyed, but I’m not going to stop considering you a friend just because my blog hit a topic that you’re not cool with. Please PLEASE focus on taking care of your well-being over reading my early morning over-caffeinated mind-dumps. So like, I’m super sorry being annoying about this, but this is something that’s getting to the point where it’s kinda making me physically ill for being treated like a binary matter.)
I am not talking about the “buzzword” version of AI.
(AGI might occur in our lifetime, but it won’t be something any large capitalist-market focused nation would allow to happen unless they go out of their way to justify digital slavery. And that is something I pray never happens.)
Nor am I going to knee-jerk condemn or praise all of ML just because of what so many people around me have been saying about it.
(Most of my working years in tech have been doing a job where the fundamental end goal is to create a tool so good that my job ceases to exist. I’ve long since made peace with the idea that collaborating to make unnecessary busy work go-away is the goal, even if that is economic self-destruction in the very real conditions of perpetual housing-fees utility-fees and nobody having enough money to fulfill all of their needs because business and governments are well aware that coercion via denial of needs is exceptionally powerful to get work done from people who would rather be doing anything else.)
Nor am I trying to claim that letting businesses suck up everyone’s data and package-that as a product for profit is inherently good, nor that processing darn near all of the data on the clearnet to train a new model for specific goals is inherently bad.
Probably one of my more genuine radical takes? Even though there’s so much fiction about how the erosion of privacy is the doom of society, I think the issue is when that erosion of privacy and control is used to reinforce dis-balanced power dynamics and to put coercive power in the hands of a few.
I think the conceptual collective ideal would be that all of this information which is supposedly “publicly available” should be treated much like one treats a library. Any kind of “products” or neural networks, etc that depend on the mass collection of all of the world’s internet knowledge, legal or otherwise, with permission or otherwise, should then be treated as forbidden to be used in a commercial context on its own, but should be fine to be derived from the same as anyone might be inspired by a book or movie or videogame or some billboard they saw 10 years ago while visiting a family member. And the underlying opposition to this matter is that copyright and IP law has long-since been used to the heavy benefit of the moneyed and only marginal value to the indie startup or solo worker. And since this long corrupt concept of copyright and IP control has been demonstrably ineffective for the mass, then it would seem most reasonable to dismantle that portion of law entirely.
People who make great music inspired by others should be free to survive for doing so. People who make music that unintentionally sounds close to something else should not have their livelihood destroyed because of over-eager automated copyright claims. The RIAA should have never have existed, whatever the well-meaning intents may have been when it started. Fanart and fandom made works have repeatedly constantly shown that having commercial control over something does not inherently come with any tendency towards actually seeking the well-being of that thing’s users nor even it’s creators, and it’s often that either from a twisted sense of legal money-seeking or from simply a lack of understanding, fandom projects that help bolster the reputations and quality of the source material get shut down or litigated out of existence or put into a state of impossibility for completion simply because the creators cannot legally ask for any income related to the thing they’re making, even though by all reasonable metrics, they absolutely have made something of quality that deserves to be acknowledged everybit as much as the source material.
We’re already living the bad-path where big business and the government agencies have defacto authority to perform this mass data gathering and processing and there are functionally no meaningful ways to enforce the eradication of the practice. So the best approach I can think of it is to have the government and businesses also be forced to have all of that information public, transparent, searchable by anyone, and SKREEEEEE. If we make EVERYONE be forced to be aware of and accepting of the reality that every single human on this earth is far more flawed than the illusion we like to project, maybe we can start encouraging more sympathy and empathy and forgiveness in the hearts of people having to witness just how vile their kin is, and that it shouldn’t make them less worthy of at least a _decent_ existence. Don’t let them sit in place hurting others, sure, but also maybe don’t run a hate mob at every single person who has followed a bad idea to its logical conclusion or someone with impulse control or someone who has poor language skills and doesn’t properly know how to say what they meant, etc
(My sincere gratitude and apologies to anyone who actually went through the effort to resize that text out of a desire to read all of it. I hope I didn’t spook you too much but I absolutely loathe the current state of IP law and I genuinely think that collective knowledge/resource pooling is the only way forward. What people typically call “style” should be sufficient branding for what people are looking for… usually. I suppose I kinda disprove my own point later on in this blog?)
In my teen years I adored Tails from the Sonic games for being super smart and making robots and mecha and stuff. And in my college years the topics I paid the most attention in were robotics, machine learning, and artificial intelligence. I still have some of my class notes from those lectures.
Well actually I just spent an entire hour checking and it looks like I kept more of the robotics notes then the ML notes. Seemingly I never took a dedicated ML class, it came up as a brief month or two of the late half of a general CompSci course.
Anyway, I got off track. What was I saying? Right, I got into the modern ML stuff. In the context of visual stuff, know how to work with depth maps, controlnets, custom LoRA training. randomized prompts, prompt weighting, etc. I’ve finetuned PonyV6 against my own drawings multiple times to see if I can get it to imitate my own round squishy “style” that people seem able to identify better than I can. So when a lot of people think of “just put words in a bar and take the image and post it” what I’m usually looking at is probably closer to Blender’s geometry nodes.
Blender Nodes workflow (Left) for generating a dynamic 3D tile field VS ComfyUI node workflow (Right, Top Portion) for making brightly colored bunnies based on stuff I’ve drew in 2013 should probably also be considered art. I definitely understand why there’s resistance to that, since the argument is a lot like the transition from paintings to photography, or from pencil sketching to digital sketching and there’s a black-box factor that didn’t exist before. But I think the core of the argument is about the same. I didn’t make every single image the neural net were based on, nor did I make every visual moment I’ve seen in my life, and nor did I make the pigments in my paint or the casing for my pencil. But I have a _lot_ of influence of what happens once the tools are in my hands, and blaming the tools instead of blaming the user’s way of handling the tools is very often counterproductive.
But what I found from putting in all that effort and time? Is. Quite simply? I even when I _do_ get the exact artwork that I want or that I had in my head, it doesn’t give me _that_ much enjoyment compared to jamming out to my favorite music tracks on my guitar, or playing a videogame and winning against a challenging boss, or even just chatting with the friends I like most and swapping idle-doodles that we mostly did to show how we’re interpreting what each-other is talking about.Things like that, any my tendency to take kind of software designer’s approach to solving issues is why I very seldom describe myself as an “artist” outside of context where friends have already prescribed that tag to me.I make things. I am a maker. I have a lot of ideas. I might be “creative”. And I do sometimes like art. But I don’t like art like an _artist_ might love it or depend on it for survival. I like art because I like the insight into how my friends are thinking and perceiving and filtering ideas and concepts that we might otherwise assume we all view in identical ways.
But maybe I’ve got a weird read on this? After all, I’m not even the best at spotting the “Revel Style” because there has been more than one time where I deliberately wanted to change identity or stay anonymous and people were able to pick out something I drew immediately, even though I felt sure it was sufficiently different than my usual.
Whatever the case. When I’m making visual stuff for Myself, doing a rough sketch as an input reference, a loose prompt, and low cfg values is probably how I’m going to get to results that I like about as much as the best of the things I’ve ever drawn for myself. But when I’m Making for myself because I want to enjoy the feel of doing something, I’ll probably stick to playing guitars or coding small digital toys.
—
Btw if you’re wondering why I’m not super hyped up about things like Udio [music] or Midjourney [images] or Runway or ChatGPT [text] or whatever the Big Budget Corporate “AI” Flavor of the week is it’s because
1: They tend to fall into the trap of thinking more scale = more better and burning an absolute buttload of energy to achieve marginal improvements instead of rethinking the underlying technology for inference.
2: With few exceptions people can’t fine-tune them on their own works to fit specific niche cases. So you’re going to end up with a ton of stuff all feeling like it was made by the same person, even though the content swap is nearly infinite.
3: I’m generally trying to achieve something _specific_ when I use these tools. I can get Meta’s Llama and fine-tune it on my personal notes and lore and run it on my _own_ hardware to be very good for _my needs_. I can pick a leitmotif that I want to carefully use across a handful of songs to carry home the idea that they’re all referencing the same narrative concept. I can come up with a specific character and re-use them an infinite number of poses or situations when I train them into the model.
4: I’m still trying to scramble my way out of a debt crunch so I’m not exactly thrilled any time one of those services asks me to pay for continued use when I could instead pay developers directly and up-front so they can continue their work while I stick to the version I already have and tweak them to my own use case.
5: I’m me. I’m aware that I’m kinda weird and quirky and have interests and vices that don’t Click for the mass market at all. I’ve said many times before that most of the things I want to make, I only want to make because I can’t find anyone making it for me or who would be willing to make it because they’re not commercially viable. So stuff that’s meant to be good at “everything” and “massively appealing” might be relevant to me sometimes, but it’s never going to
okay I guess I spent like 3 hours pouring out my thoughts instead of just the few minutes of “trying to explain what I mean by Making For (xyz)” and I need to go earn my paycheck from this dayjob so uhhh yeet?