Three years ago, I was trying to figure out what to do with my life. I had some savings, and around this time, AI was hitting the mainstream.
DALL·E was obviously astonishing, but what made me fall in love with AI art was what happened in the open-source community after Stable Diffusion was released. Every day, like many, I'd go on Reddit and see people extending these models in all kinds of unexpected and ingenious ways.
It felt like looking through a window at something potentially utopian: people tinkering with cutting-edge models on their computers and sharing their results for others to learn from and use - all for the purpose of creating strange and beautiful art.
For months, I barely slept, trying to get a grip on how this worked - training, prompting, fine-tuning, everything.
I quickly realised I was helpless on my own. Thankfully, there were lots of others like me, on similar quests.
So I started to build a community of people who were interested in doing meaningful things with this tech - actually making art, creating new tools, and exploring it seriously.
That became the Banodoco discord. There was a lot of attention around the degenerate uses of this technology, but I wanted to create a space dedicated to the more wholesome, open-spirited part of it.
And it worked!
For the past few years, we've built what's become a beautiful space where talented people create and hang out together. I personally find it very inspiring.
Since then, many of us have been working on and off to push things forward - building technology, making art, sharing generations, helping other people - and mostly doing it for nothing.
That all sounds pretty utopian, right?
But if it were, I wouldn't be deeply in debt and I probably wouldn't be writing this post.
On the face of it, everything seems fine - great new models, more tools, better capabilities.
But as someone who cares deeply about the heart of this - the thing I fell in love with - I feel a growing sense of anxiety about where it's heading.
The beautiful, open ethos that initially inspired me is still there. But I worry it's being hollowed out.
Partly, it's economic. Many people who care about open-source models have no real way to sustain their work. Some drop out, some burn out, and others become demotivated.
Partly, it's aesthetic. As soon as these models could create things that look very realistic, the center of gravity shifted there. A lot of the work got less novel and less interesting - often just replicating Hollywood or chasing the most technically impressive thing.
Partly, it's cultural. The main public spaces for this stuff have become overwhelmingly dominated by porn, to the detriment of art - to the point that the largest one has been banned in multiple countries.
But a big part of what saddens me is the wasted potential. So many models and directions that are unexplored and unexploited - so many opportunities to create remarkable things that never take place because people get distracted, they're busy, or they lack the compute.
Finally, as these skills became professionally valuable, what started as playful experimentation became work. A lot of talented people understandably got pulled into the closed commercial world.
And every time I see a company close its models, or someone doing meaningful open-source work give up, I can't help shake the feeling that if the broader ecosystem were more alive, more vibrant, maybe those decisions would be different.
The fire hasn't gone out. But I feel this anxiety that it's fading.
While working on countless open-source-related projects, my goal has been to build an open-source-native company, Banodoco, that helps build tools to expand the open-source ecosystem. Because I wanted to "go all the way" with this, I decided to give 100% of the company's ownership away to people in the community based on their contributions.
I still think this is the right thing to do - especially given my goal was to build a company on top of their work. But unfortunately, from the perspective of many investors I spoke to, it makes the company "uninvestable".
I had a bunch of conversations with people who were interested in investing, but only if I'd make the structure more conventional. And I couldn't do that. Not only because I'd made a commitment, but because I knew it would be to accept a compromised version of what I felt was needed.
So instead I struggled on without meaningful resources, without much ability to act on what I believe beyond my own energy and time. I'm still not sure if that was the right decision.
All this to say, when I woke up the other morning and discovered that Elon had retweeted a project of mine, and that some crypto nerds had created a meme coin in its name - people gamble on these things, and a cut of every trade gets paid to the creator in Solana - I felt like I should try to do something with this somewhat "dirty" money. I'd accumulated over $15,000 in these fees!
So without thinking too much about it, I tweeted that all of it was going to an art competition focused on open-source AI art.
This caused a lot of speculative gambling on the coin. A lot of people told me I can't just give it away to something unrelated. But all that resulted in was more gambling. This caused the fees I collected to quadruple to over $60,000 in liquid Solana that could be easily converted into real currency.
Today in particular, I feel as though this might have been a terrible decision. $60,000 is, coincidentally, around the amount of debt I've accumulated after eating through my savings. I feel like I've been throwing good money after bad, and that I should have just packed it up and considered this freak act of luck to have neutralised my losses.
Firstly, this money and competition probably won't have a meaningful impact. A few dozen people will make stuff, a few hundred will look at that work, and this time next month it will have done nothing.
Secondly, I feel as though people just don't care as much as I do - or for various reasons don't want to support it. For example, I posted about the competition on the main subreddit for AI art - one whose moderation lets it fill up with soft porn! - and the post got removed, as it has every time I've posted about previous iterations of this competition.
It makes me wonder: what am I doing with my life? Why would people not only not want to support it, but actively want to diminish it? Why am I doing this? Who am I doing it for?
But while I felt that at the time, it doesn't fully capture how I feel.
For example, every day I go into the Banodoco Discord and I feel the warmth of that flame I fell in love with. It's still burning there.
And while I have no idea if the competition will be a big success, I get a bunch of messages from artists whose work I love - often people I haven't seen anything from in as much as a year - telling me they're making something for it.
And while many who contribute here have given up, I spoke to a friend this morning who's been doing this work for years. He's still doing things, still full of ideas, and still pushing it forward.
I've also seen people in the community making remarkable things that others can use for their projects - LoRAs, tools, resources. I put some other crypto money I got towards Art Compute, which gives people in the community compute to train models - and I see talented people already doing impressive work with that.
I also see people like Zeev, the CEO of Lightricks, committing to releasing more open models. Lightricks are also supporting our upcoming event, and Comfy.org are also supporting the competition.
And despite despising crypto for a long time, if I really squint, I can see some way in which crypto can fund open-source work and create an economic outlet for people and projects that don't fit conventional capital structures.
So, to answer the question I posed in the headline: yes, it was extremely stupid for me to put this money towards this. I probably should care less. I should probably just get a normal job - my mother certainly thinks so.
But despite how I woke up feeling this morning, and what some might see as an objectively dire situation, that isn't what I'm choosing today.
Next week, I'm sharing a tool I've built to help bring the artistic power of open models to more people. After that, entries close in the competition and we'll have a public vote. Then I'll launch a new space where people can share resources and art made with open models - one that isn't infested with porn! And next month, together with our friends at Lightricks, we're having an event in Paris to bring together more people who care about this.
Yes, giving this money away is probably still stupid. But maybe it'll catalyse people into creating real art. Maybe that art will touch someone. Maybe it'll reach people who've been thinking about giving up.
Or maybe it'll do none of these things. Maybe it's just plain stupid.
Images generated using Matteo's IP Adapter for ComfyUI with SDXL