Welcome to 2024 peeps, where tech meets art with innovations so wild, even your toaster might be a part-time abstract painter. The realm of AI art is a fascinating frontier pushing the boundaries of what we once thought possible in the world of creativity. In this conversation, we have the pleasure of delving into the mind of Picatrix Picori, a renowned AI artist known for his captivating manga-style illustrations.
Background:
Picatrix Picori's journey as an AI artist specializing in manga-style illustrations is nothing short of remarkable. His passion for this unique art form was ignited during a pivotal moment in 2012 when he embarked on a quest to explore computer graphics, physics simulations, and semantics, leading to an extraordinary fusion of his musical talents and radical obsession with technology.
Inspiration and Creative Process:
Drawing inspiration from the likes of Joseph Campbell, Alan Watts, Ken Wilber, and Robert Anton Wilson, Picatrix masterfully crafts AI-generated art that delves into the depths of the human psyche while delivering aesthetically stunning results. What sets him apart is his willingness to push the boundaries of AI artistry, creating pieces that might be provocative or even offensive in pursuit of preserving the non-corporate persona of AI art.
Tools and Techniques:
Picatrix employs a variety of AI tools like Automatic1111, Playground, Bing, Starry AI, Unstable Diffusion, and more to create his manga-style illustrations. His meticulous attention to detail and quest for a handcrafted, authentic feel set his work apart in the digital world.
Challenges and Vision:
As an AI artist, Picatrix has faced challenges in carving out his niche and defining his style. He believes that balance between maintaining artistic integrity and adapting to technological advancements is a constant tightrope walk. His vision for the future of AI art is a decentralized movement free from corporate control and beholden to the principles of artistic expression and individuality.
So, let’s kick off our the interview with Picatrix Picori!
IG: picatrix.picori
Can you share your journey as an AI artist specializing in manga-style illustrations? What inspired your focus on this unique art form?
Picatrix: In 2012, I left college and found myself feeling like an outsider in a corporate-dominated world. I became obsessed with learning about computer graphics, physics simulations, and semantics. While I initially pursued music and tried to fit into the "Boiler room" culture, I didn't like changing myself to belong to a culture that didn't prioritize people like me. Instead, I focused on radical virtual ideas, escaping reality, and prosthetics because I believed technology would lead to a different direction than the outdoorsy culture promoted by movements like Algorave.
When lockdowns happened, AI art gained popularity, and my ideas about isolationism, virtual worlds, and graphic enhancement became more relevant. The convergence of these ideas with Philip Rosedale's involvement in Midjourney felt like a dream. I had already been playing Second Life and anticipated the Wachowski sisters' timeline becoming more relevant. This convergence seemed unlikely a decade ago.
Can you describe your creative process, from ideation to completion, and how it differs from traditional artists?
Picatrix: My approach relies heavily on fairy tales, narrative tropes, and character development. My work relies heavily on fairy-tales index and narrative tropes to understand the different “crisis” in a character’s psyche, also to understand how the character can learn from problems and develop as a sentient being. I draw from Joseph Campbell's monomyth and combine it with a more materialist and deterministic viewpoint influenced by Alan Watts, Ken Wilber, and Robert Anton Wilson. Joseph Campbell’s monomyth was cool during the Psychonaut craze of the previous decade with Erowid, Bluelight and other forums full of people talking about their use of psychedelics in the light of Arthurian Folklore.
When writing a prompt for my AI art, I aim to summarize meanings in a surreal but semantically rich way as it is a cluster or overlap of images but I approach things from the point of view of semantic categorization to make each word enhance each other and achieving better aesthetics with the average AI tools. I leverage motifs, stereotypes, and statistical prevalence in the training dataset to enhance aesthetics, but also as an opportunity to twist the plot with funny AI generated bloopers. AI often interprets things in amusing ways, resulting in surreal combinations, such as a horse's leg overlapping with an ice cream. Although it wasn’t such a good idea to generate Rasputin as a three-legged chair lol.
AI technology has revolutionized the art world in numerous ways. How does your work in AI-generated art differ and stand out?
Picatrix: My art might offend some people because I've explored the boundaries of AI art. I've offered my prompts to scanlators and translators for creating Shojo Manga pages, but they rejected the idea, fearing it would make them full-fledged mangakas. My focus has been on addressing the fears surrounding AI art and pushing the limits of AI art without succumbing to corporate pressures or the NFT hype.
I dislike the idea of AI art becoming part of the Bored Ape Yacht Club or the crypto-culture. I was pleased that the manga scanlators rejected the notion of enhancing their public personas, but I was also somewhat disappointed that this group chose to reject AI tools and allow newcomers to disrupt the manga industry. This mirrors the typical clash between established brands and imitators in the AI art world.
Corporate competition shouldn't be limited to an exclusive group. I want to share my prompts with dedicated artists who aren't driven by financial ambition, creating a DIY scene akin to Garage Rock/Punk, free from overpriced prompts controlled by old companies responsible for the music industry's hypocrisy.
Like Al Jourgensen, the founder of Ministry, I believe our primary focus should be on creating quality art without corporate constraints or editorial obligations. AI has the potential to eliminate jobs, but I won't sell this "secret" to corporations for profit. Instead, I want to provide it to people who appreciate and connect with authentic art that empowers them to be themselves without conforming.
How do you balance capturing the traditional manga aesthetic and storytelling elements while maintaining the AI aspect of your creations?
Picatrix: Interesting! Because I know this is important to be addressed. When people who are rather misinformed about AI manifest their opposition and fear they think we are gonna be some sort of replicant doppelgangers impersonating their favorite idols or making nude pictures of their daughters. In Second Life there is a word for this behavior, ‘griefer’ is a person who just does this and has zero value as part of a support network to make people feel autonomous, a ‘griefer’ is a cybercriminal and AI Artists are not -generally speaking- like these.
So, we can not just rip-off mangakas and pretend to be the new Akira Toriyama like a crypto gold-rush. Beeple is very humble despite being the go-to example of what this trend could be like. Emad Mostaque had his own problems with the way brands felt affected and I would not like to do that, even if there were no lawsuits I’m broke anyway and I can’t buy branded products enough to care about them so I just make my thing.
I like to consider the Doujinshi laws about nudity and try to come up with “universal” guidelines of censorship for my own work, also to respect brands and be open about fan-art being just fan-art not some product I would be entitled to distribute and profit from. This is got a lot of influence in the way my art looks and these limitations helped me create a style… which I don’t think is super popular yet but the European legal system will eventually push other AI Artists to think this way as more and more AI Artists come up with ways to serialize the production of AI Generated commodities that could impact the market like a corporation could… and without breaking copyright laws! Which is not impossible.
Can you explain the tools and techniques you use to create manga-style illustrations, and how these tools have evolved over time?
Picatrix:I use various tools, including Automatic1111, Playground, Bing, Starry AI, and Unstable Diffusion. However, Dalle3 exploited its corporate infrastructure to offer cheaper services, which pushed me to develop a more resource-efficient approach. I’m not afraid of the word cheap or minimal consumption since I think it is a bit of a mind game to come up with a way to achieve technical proficiency with minimal resources. The lack of version control for AI art tools can be frustrating, as sudden modifications can disrupt the artist's workflow. This should improve as the industry matures.
New tools could eventually make automated text generation, use a story index to have templates of scenes, color combinations and stuff but that is not really necessary unless you need an army of bots to invade every corner of the internet.
Upscalers are crucial for creating posters and interventions in the urban landscape. Oracle introduced a blockchain-based resource mining framework, and there are numerous prompt-sharing platforms. Still, the hardware constraints necessitate bargaining with the GPU limitations.
Probably my comfort zone is de-centralization groups and when I think of tools I think of empowering this production assembly for decentralized end-users and long-tail market as a way to avoid being co-opted into these funnel metrics that just make a puppet of every influencer with more than 2k followers. Co-opting elites of small communities into the “Valhalla” of influencer stardom is wacky and I like to research ways to avoid this from happening.
So, tools for AI Art can not become a monkey on the back for people to become bitter and desperate when they don’t have them, which was a little bit of a problem with Colab Notebooks in the early days of Automatic1111 Webui.
We need to make better technology to circumvent the GPU hegemony and other hegemonic suppliance constraints. CUDA, Colab, GPUs… luckily Oracle came up with a framework for resource-mining using blockchain and there are tons of prompt-shares!
Although being honest, the initial sense of gatekeeping and “Da Vinci Code” conspiracy is slowly disappearing just like the Anti-AI mobs kind of disappeared (Nobody has coded a virus to steal CPU resources as a revenge against cloud-services, right?)
Just like Emad Mostaque and services allowing free daily credits, a lot of these major players like Microsoft have been very kind at making these tools available for people, Cloud-based GPU is a nice development… but still the hardware constraints force us to bargain with the devil, there are standard procedural guidelines that have been ignored to make a minimum viable product very fast and the lack of version control to access weights that were functional in the past becomes an accidental tool for oppression when a “magical prompt” stops working all of a sudden… because of fine-tuning that affects the end-user and beta tester in non-consensual ways or without previous notification, that affects the end-user’s own scientific itinerary of semi-structured trial and error so sometimes is hard for students and researchers to have control of this “magic eight ball” generating random pictures.
I know what theories about flux and relativism and VUCA-BANI say because I am a Product Manager as well, but for a computer to have a stable performance they shouldn’t mimic in their programming how f*cked up society is. Ignoring standard procedural guidelines gives a sense of relativism and praise for sudden modification that could give the wrong impression to cyberterrorists sabotaging AI. This sense of outlaw manufacturing is probably going to disappear if a proper business model for commercialization is devised, yet the completion of this industrial standard would also mean the destruction of the emblematic protagonism of hand-made paintings in the history of mankind; personally I’m not afraid to make prompts that solve some of these issues and relocate or redistribute art’s motherland out of California and Paris.
More control over prompt determinism isn’t the same as prompt weight or denoising but I think we kind of achieve it using ControlNet, LORA and img2img when using Automatic1111. The output should be more “consistently candid” or so they say.
Your work seems to encompass both traditional manga elements and modern AI technology. How do you strike a balance between these two worlds in your art?
When I was a kid I liked girly manga and very technological and dark mangas, just those two styles; traditional gender roles were a little bit frustrating. And the same happened with literature genres and art genres. Being a mongrelized Asturian African-American Jewish person I felt these genres were very square with the whole boy-meets-girl structure of the stories that typically appeal to most latinos, however I liked the shyness of the most conservative and naïve characters of Shojo manga.
As I wasn’t part of a tribal structure where I could naturally “explore and grab” like most people did during the haydays of Reggaeton, I didn’t like the stories with these romances that were too easy to predict in a timeline from point A to point B. So, while avoiding these “hook-up” narratives, techno stories of isolationism and Shojo stories of shyness really caught my eye because they felt more elegant and aware of the technological future of mankind.
I think Hatsune Miku is a good example of a girly form of futurism and how this intercultural impossibility for hook-ups made a whole industry of artificial experiences. So, when I make my art, I feel naturally benevolent towards readers in a similar quest, even if I’m straight I feel a lot of empathy for transgenders in the e-girl world and online virtual worlds but also for people who express their rarity in a macabre yet beautiful loneliness. This rarity doesn’t really reflect in the rarity that inflates the prices of NFTs, that axiom influences my work.
The bishonen characters of some Shojo manga make me laugh, I have little respect or interest in those because they seem a little bit too squared and submerged in their whole hook-up roadmap which is like a straight line with little crossroads unless is cheating their girlfriends or some misunderstanding between the two, which is not easy at all to portray with AI Art that for now relies a lot in elliptical sequences to narrate stories by using little frames. I think that when discussing tools to make my own craft, these lifestyles that I relate very little to have coincidentally had equally little impact in the availability of tools made by people in these niches, this is another statistical axiom that I consider when choosing demography and style and also being inspired by an idealistic notion of charity and affection.
So besides the very shy demographic agenda to make new forms of pluricultural and subcultures in a society that, despite the harsh inequality worldwide, has reached indicators of a possible post-scarcity future there is a Product Management criteria to consider of what is possible to represent using AI and what really gets me in the mood to write stories… sometimes the story writes itself and the picture requires just some minor scores to achieve a proper presentation and format, is a lot of fun and i hope the leap of faith to dive in my own uniqueness is rewarding both in human, financial and professional terms.
What challenges have you faced as an AI artist, particularly in creating detailed manga-style illustrations, and how have you overcome them?
Picatrix: I remember one year ago during Christmas 2022 when we met, you (Akasha) required some concept designer for the characters of your short stories. I realized you were approaching things in a way that makes me think in the previous question, I could notice your story’s timeline wasn’t exactly Shojo but you had approached character design with a bishonen aesthetic of which I was very curious because yaoi manga is generally seen as a rebellious stomp to the heterosexual “hook up” narratives and it plays with disillusion a lot when gay characters prefer other men instead of the woman that is in love with them, which often is like Sophie the lead character of Howl’s Moving Castle but full of fanservice like Visual Kei.
At first, I didn’t know which trend to jump-in to. So, one challenge was to pick an editorial line and put on hold the desire to revolutionize aesthetics with the power of AI. I hope our results have matured well with new developments of AI like Gorillaz the rock band did with the downfall of “Laguna Beach”. The client being part of the real-time selection process to fine-tune the result helped me very much to meet stakeholder requirements.
One challenge I am very proud of having solved is pixelation. Early versions of Stable Diffusion had a way to paint using noise that if you zoom these pictures you will notice the trace is not a hand-made stroke that projects a color frequency at a distance, so i had to make use of my limited knowledge of physics simulation and make a physically-based approach to prompts. If you zoom in on my manga work with AI it really looks like it was hand-made with pencil or graphite powder and that is because I researched laser printing to simulate burned dust on paper.
When Camenduru’s webui spaces got rid of the ‘extras’ tab with which i was upscaling i decided to do a little trick. My new pictures don't require upscaling because of this but the whole aesthetic is really taking me away from the Midjourney hyper-realistic trends of digital render of the AI Art Community and this sort of “Madrid fashion week” aesthetic. This limits the expression of my prompts but also makes the art have a branded personality that i wanna put to test by giving away the prompt as a gift for New Year, if i succeed i will feel as cool as Emad Mostaque, lol.
Another challenge was to achieve independence between the semantic coherence of the prompt applied to the setting and the one applied to the character. It was too symbiotic when making a concept the character had to fit the rest of the prompt and vice versa. Dalle3 fixed this for its Bing image generator and that’s a lot of expectation regarding how Bing will evolve cus it seems to be the only model that so far allows to manipulate the character, the pose, an extra element and the setting with separate sections in the same prompt and for the same picture.
[Here’s a link to the book Picatrix and I collaborated on, generating AI art to visualise the characters - a LOT has changed in AI art in 12 months!!!!]
How do you stay inspired and continue to grow in a rapidly evolving field like AI art? What resources or communities do you rely on for support and inspiration?
Picatrix: I used to be inspired by cyberpunk, and cyberpunk was my lens to scrutinize the possibilities for modification and upgrade of my work. But, as those “stranded hippie” years fade away and I’m becoming more fit for a industrialized structure in the new tech-startup landscape; because of superior education that I've certified during lockdown I don’t think like Deltron 3030 anymore.
At first it was a primitive need for survival but now I feel more comfortable taking a lead and risking myself for an incremental welfare or even changing the world without profit but really making sure the change indeed happens.
Communities like the AI Art Community, Second Life, and Print on Demand have provided opportunities to learn and connect with others. I like being incognito among those and learning what I can, especially when they meet a common ground in decentralized interests and I can join a streaming and do benchmarking. I've been an early adopter of the "pirate-entrepreneur" mentality, focusing on producing art with minimal consumption and avoiding corporate constraints.
S
What do you predict for the role of AI artists in the future of work as AI integration continues? How might the industry evolve?
Picatrix: I foresee a rise in "pirate-entrepreneurs," artists who create independently and prioritize autonomy over corporate involvement. I mean, people who have enough studio tools to really dive in their craft and worry less and less about a support network to maintain a relative welfare that allows them to continue production with minimal consumption of resources and basically zero reinvestment. Probably a bartender at Starbucks at day, a prompt engineer by night is the kind of cool dude of the future; not some influencer jumping to the celebrity system and overdosing like Ziggy Stardust or becoming some snitch for a shady cartel.
The AI art landscape may shift away from the influencer model, and artists could focus on developing their unique styles and brands. Although this has never been my priority and it's just the natural trend of technological upgrade, I think the global domino effect will make us become more like Gen Z, in some ways non-dependent to advertising or traditional metrics of success in digital marketing. My concern is the audience that is too attached to a societal model of “overseers and maroons'', which I shall not call “the” since there is not a single audience, there are tons of audiences and radical politics demonstrate that crazy ideas of hate and exclusion do great and thrive in the market when an opposing demography embraces them. So, these pirate-entrepreneurs will be “underground” and have zero participation in the political game of blaming the other to capitalize through polarization; since politics and mainstream entertainment became like “scareware” tactics.
Underground is better and I've been trying to be an early adopter of this pirate-entrepreneurship which should not be confused with fraud by making pirated versions. We need leadership and community-regulated guidelines to avoid copyright infringement with the proper sourcing so we can finally free ourselves from editorial obligations typical of old business models based on advertising, investors and sponsors. We should be like people who record, mix and master their own songs so they can say whatever they want as long as it isn't too crude. In this way Philip Rosedale is a very benevolent presence in the AI Art world because he really achieved this mature attitude in his clients and players of Second Life.
In this way heterosexual bishonen, metrosexuality and self-promotion make a chimera of joint forces against free market and against “underground” counterculture that is being empowered to pose as science/tech gurus, which is very dangerous in the context of disinformation.
This self-promotion hunger began with Dj’s,Tenacious D and Guitar Hero twenty years ago, let’s say you were with your friends watching YouTube and you played a song cus you thought you were enough of a deejay and had been practicing, then suddenly everybody was taking turns to play a song each cus they too were convinced they were deejays… but none had ever spun a record or even used Traktor they just imitated and enforced an a priori right to exhibit themselves. This happens, among other things, because of the availability of technology for people to fake it til they make it and artificial intelligence is another niche of the decadentism fever dream of artificial paradise proposed by Baudelaire that inspired cyberpunk and dandy culture.
In my city a term for this process has emerged but its not really a likable word even though it became mainstream urban iconography, it is “espanta jopo” or “ass-scare” and its the way local Caribbean-Colombians understand posers from cultures they themselves don't belong to but are kind of funny to scrutinize because they stray far from the autochthonous idiosyncrasy and experiment misguided transformation that is undesirable for sexy women, which is why they “scare asses away”. It's got a lot of implicit macho-culture but it has certainly identified a social phenomenon a lot of people can point at and hold a shared opinion about it. It could also be said “booty-scare” but the Spanish word ‘jopo’ is more of a creole word with a very crude connotation.
This is something that in latino communities makes people earn rejection and the spokesperson for the scanlators i approached with my “magical prompt” was probably concerned about being an overnite prodigy, like that video for ‘Windowlicker’ by Aphex Twin, that’s probably something macho-centric cultures still relate to everything digitally crafted.
In the future, AI artists may need to choose their authorities carefully, as community-regulated guidelines gain importance. It's essential to navigate this evolving landscape with a focus on creativity, authorship, and the development of a personal mythos.
The whole funnel metrics based on views and traffic will disappear eventually and these “savant” consumers will probably burn their desire with infinite generations and eventually re-focus on their family life and start supporting dedicated artists without the selfish motivation to amplify themselves.
With AI Art and its 1 year old developing culture it's already happening because this behavior became a de facto presence in the social landscape (Twitter?) and there are memes making fun of startups making API calls and new-born CEOs, but this is the least likely burden avant garde artists will suffer. There is a lot of freedom but it’s not like a Fender Guitar you can buy and do whatever you want with it without disturbing the manufacturer, in AI Art there are overseers, some like Dolores Umbridge and some like Dumbledore, even from outside the circle of authorities in AI Art, and I hope one day this changes when community-regulated guidelines are stronger.
Long term fulfillment of certain conditions can be speculated but Its soon to know how will the AI Art World choose its authorities, besides Elon Musk who’s allowing us to multiply and thrive in his social network (even though a lot of artists blamed him of shadow-banning cus they didn’t read X’s new interaction guidelines about influencer cartels before jumping on the Threads and Mastodon wagon) we’ve got people we respect like Camenduru who’s very punctual and decent, Emad Mostaque who is got his own charm for those who like his style, and that “senpai” blend of company spokesperson/community manager who’s on social networks more oriented to the “swag” lifestyle of AI Art and NFTs and will happily like your AI Artworks and retweet (don’t tell Elon I said retweet).
Probably the most respectful artists are those making artistic nudes and a small community of waifu artists (from Japan?) that really makes some beautiful art that differs a lot from average anime fanart and furries that are generally considered as a creepy niche even in Second Life. There is a lot to try and awesome possibilities to dive and really develop one’s talent and brand.
What advice would you give to aspiring AI artists interested in making art?
Picatrix: For new AI artists, I encourage them to embrace this new culture and methodology while leaving behind the "meme-world." Build your unique universe and focus on developing your skills and brand. Avoid the "fake it till you make it" mentality and aim for long-term fulfillment and growth as both humans and artists.
When it comes to AI Artists trying to be influencers, I personally didn’t enjoy parodies and this phase of AI Art was too easy to relate to all the lawsuits and existential dilemmas like the “No-AI” movement which was a little bit violent and for a while AI Artists were like little scared siblings protecting each other from the fanatic mobs from ArtStation.
So, a good idea would be to develop creativity and authorship because so far there have been no issues with people making their own universe, for this reason im sharing a very reliable prompt to make Shojo Manga which the user can customize to get his own style and shapes, and luckily make an underground magazine or a punk fanzine dedicated to Ai Yazawa.
Have a happy 2024!
:::::NEW YEAR GIFT - PROMPTS:::::
Add these prompts for people to make Manga with Dalle3 using Bing's image generator.
If you made it down this far, thanks for reading!
In the radical landscape of AI art, Picatrix Picori stands as a pioneer of the "pirate-entrepreneur" approach, emphasizing the creation of art with minimal resource consumption and a detachment from corporate constraints. This shift away from traditional metrics focused on views and traffic heralds a future where artists are supported for their unique contributions rather than their ability to self-promote. Picatrix's vision suggests a future dominated by independent artists who prioritize creativity and personal mythos over mainstream influence, navigating a world where community-regulated guidelines gain prominence. This evolving landscape presents an exciting opportunity for artists to develop their unique styles and brands, redefining the essence of being an AI artist in the digital era.
Like what you’re reading? Come and read my blog and listen to my interviews over at www.futureofwerk.xyz