Seeds of Creation
There is a plant native to Madagascar, the native peoples call it 'Songosongo', when the species was introduced to France it took on the name: 'Crown of Thorns'. This plant is both beautiful and imposing at the same time. Lush, soft, leafy green stems, dotted with crimson flowers, spring forth from rigid, strong branches covered in some of the most intimidating thorns seen in nature. The texture of the branches is dense and woody, covered in long sharp thorns that look incredibly painful. These evoke the crown of thorns Christ was made to wear when crucified. This plant signals clearly: it should be approached with caution.
In my last article, I threw a stone in the pond. My intention was to try to convince you that we need to stop exerting ourselves on throwing stones and instead focus on the ripples we can create. I promised I'd detail some real ways we can confront our true enemy, and use our creative energy to build a future where we can flourish. A future where our individual intellects are expressed, worn as a floral wreath. That these words act as seeds that will grow into a crown of thorns, meant to protect the blossoms of creation.
AI is not as much a threat as we've been made to think it is, at least as it pertains to art. Yes, it will fundamentally change our economies. To me, it's like an advanced extension of clip art. Albeit, extremely advanced. I think it's well suited for marketing tasks, which let's face it, don't really require authenticity. I once did an internship where I made animated web banners... While the work was creative, it's kind of a stretch to call it artistic expression. Yes, jobs that encompass this kind of creative work will largely go extinct. We intrinsically value AI content less than human created content. AI-generated stock photos are already priced at a 40-70% discount to human-created equivalents. A 2023 study found consumers could only distinguish AI-generated text from human text 50-55% of the time (chance level), BUT when told content was AI-generated, they rated it significantly lower — the label matters more than detectable quality.
The fact is, it needs to consume more and more original content, to stay relevant. What does this mean? It means, it needs you more than you need it, at least at this current juncture. Picture the Ouroboros, but with each revolution the circle becomes smaller, as the snake consumes more and more of its own body. Research published in Nature (Volume 631, pp. 755-759). This landmark study demonstrated that large language models, variational autoencoders, and Gaussian mixture models all degrade when successive generations train on content produced by earlier models. The model's view of reality narrows, rare events vanish first, and outputs drift toward bland central tendencies. The researchers termed this "model collapse" — indiscriminate use of model-generated content in training causes irreversible defects in resulting models, where the tails of the original content distribution disappear.
This is where the first solution comes into fruition. Clearly at this point, within online systems, we've made ourselves into low-hanging fruit. You see the current web paradigm, people give away massive amounts of their original content to a gamed, predatory system. Increasingly, this has become a losing game. Just take a look at these stats:
The Web2 content economy is defined by extreme asymmetry: Meta's 2023 advertising revenue was $131.9 billion; YouTube's was $31.5 billion; TikTok's was approximately $16 billion. Of this combined $179+ billion in advertising value generated by creator content, only $16-20 billion flows back to creators — roughly 9-11% of the value their content generates. Meanwhile, revenue concentration is extreme: the top 1% of creators earn 80-93% of all platform revenue (Stripe 2021). The median Patreon creator earns ~$50-100/month. Only 12% of full-time creators earn more than $50,000/year (Linktree 2022). 46% earn less than $1,000/year. The paradox: a $250 billion creator economy with 200+ million participants where fewer than 1% earn a living wage (Goldman Sachs, 2023; SignalFire).
By harnessing the power of this apex predator for systems creation, we can make the corporate monolithic gardens wither away. There is nothing particularly innovative about most centralized tech platforms. In fact, it could be argued that the experiences continue to degrade as the surveillance capitalism economy advances. This is because the model is flawed, and must become increasingly more intrusive in order to stay profitable. This is why these monoliths are destined to wither away and die.
This is why I reiterate, we are looking at this whole thing in the wrong way. Projects like Freenet and others offer the promise of completely decentralized online systems. We can build these systems in ways that benefit artists and creators from the outset. We know now how malevolent forces have tried to use these outlets for control. Our blossoms of creation are easy pickings on every online platform. We need to take a cue from the Sosongo, grow thorns to prtoect every blossom. Truly, we can make their efforts all for naught. What would you rather do? Post on Twitter for the likes, likes that are not in any way organic. Argue with an endless ocean of bots that exist on these platforms. Or would you rather dig them out by the roots, and watch them burn? Fertilizing the way for our own fields.