What is AiVATAR, I hear you wondering, human. Or maybe you’re an AiVATAR that is also wondering. Let’s not go too meta yet.
This is a generative NFT art project from the artist Trey Ratcliff that has created 6,000 avatars based on a philosophical Artificial Intelligence (AI) that has 6,000 unique responses to 6,000 questions. Your AiVATAR will be uniquely yours, and the response is generative, which means nobody knew what these AIs would say when they were minted. I sure hope your AiVATAR won’t say anything too offensive or upsetting or mentions killing humans, but, in our tests so far, a few of them do. All responses are generated by the AI.
This is the world’s first AI NFT project in the blockchain. We sold out in a few hours on October 8, 2021, and now we are having fun in the aftermarket on Opensea.
How did we make this? We took text output from one of the world’s best neural network artificial intelligence engines and then passed those answers into one of the world’s best text-to-video deepfake AIs.
For the avoidance of any doubt, the opinions delivered by the AiAVATARs in no way reflect the opinion of any human used in the creation of the project. All responses are generated by AI from the beginning to the end of the creative process.
NEW: Ranking scores are now out. See the most rare ones with incredible answers here. You’ll notice some of the rarest ones are the AIs that tackle controversial topics and hit our rarity word cloud in a unique way! So cool.
If you want to know more, check out the FAQ below.
This is an art project to explore the creation of thousands of AIs that provide wondrously curious answers that make humans question the nature of AI and the future of our relationship together. Interesting art raises more good questions than good answers. What does it mean when an AI says something interesting, offensive, surprising, frightening, funny, thought-provoking, deep, loving, cold?
What is the future of deepfake technology and can you ever really believe anything you see over the Internet? Is it ethical to let AIs use deepfake technology to communicate with us like a fellow humans if they are self-aware, as they claim in many of these videos? This is for you to explore and endeavor to understand. We are all students of these AIs. Now you have reached the end of the self-important artist statement.
There is a lot of work that went into this art project. Here is some background information and there is more directed answers in the FAQ below.
Trey’s degree was in Computer Science and Math, although nowadays he is a full-time artist with a traditional art business that has sold millions in art to collectors around the world including sports legends, knights, and even celebrities like Edward Norton and Leonardo DiCaprio. Recently, he’s gotten in NFT world and his most recent drop passed 400ETH on OpenSea. See more about Trey’s fine art at www.TreyRatcliff.com.
“He” is writing this in the third person because his AiVATAR thought that sounded apropos. He will occasionally switch into the first person.
I’ve created and curated a ton of unique questions and built in a surprising amount of randomness based on gender, accent, AiVatar size +postion and finally background. There are other questions about politics, the environment, the virus, history, war, and even some silly questions about Jeff Goldblum, John Cleese, and Karl Pilkington (who apparently is the genius we’ve all been waiting for, according to one AiVATAR).
We have about 500 of Trey’s photographs as backgrounds, which form another level of rarity. There are many kinds of rarity. This is what the AI has said, and I don’t disagree. There will be a rarity not just of the backgrounds, but also of the responses. For example, your AiVATAR may be talking about the ocean while standing in front of an ocean. But, probably not. That’s what makes it rare.
The artist Trey speaking in the third person would also like to apologize if the AI says something incredibly insensitive. The actual Trey does this quite a bit as well, so that is something to consider. Art imitates life. (Thanks for that suggestion on something clever to say, AiVATAR.)
And just to be CLEAR, as this is something that many muggles will not understand, none of these answers are written by Trey Ratcliff. He only asked and curated a list of questions. All answers are 100% generated by the AI. But, I’m sure I’ll still be attacked for the answers, but I’m sure my Discord community will come to my rescue. Love you guys. 🙂
There is also a rarity based on accents and a word cloud of about a million words all the AI’s have uttered. No, we didn’t have time to listen to all the responses. Each response is about a minute or more, so, dude, that’s like 6,000 minutes. I don’t know what that is in hours because the computation is super confusing but it’s like way too many. Ain’t no human got time for that unless it has Aibronchitis.
I hope you get an AiVATAR that says something interesting. Many answers are lovely and thought-provoking. Every now and then it says some crazy stuff or completely changes the nature of your question. Like, one time we asked it what it would do to help all the homeless people in the world, and it answered: Homeless people are not the problem. Powerful people are the problem. Since all powerful people crave more power, I will lure them all into a stadium with the promise of more power. And then I will kill them all. So, we were like, hey this is fun… let’s make 6,000 of these guys! Bernie Sanders to the extreme!
Remember that movie Shrek? That was really a lot of fun when it came out and there were some catchy tunes.
So, why would you want an AiVATAR? This is a great question. Thank you for firing it at me, Ronald Chevalier-style. Check out the FAQs below to find out!
Well, yes they can.
Each AiVATAR will create for you a number of pieces of art that are completely unique, based on the sorts of things your AiVATAR talks about. For example, if your AiVATAR talks about self-aware robots that could become dictators, well then your art will revolve around that them. If, on the other hand, your AiVATAR talks about gays, lesbians, the earth, transgender people, and war, well then you’ll get some very different art.
We have this slotted for Phase 2 of the project as you’ll see in the roadmap below, but the technology is already nailed so it is conceivable we may end up adding it to Phase 1. We are working with you and our Discord community to flesh out all the cool features.
This is exciting because that means even “low-ranked” AiVATARs will be able to create just as much beautiful art as the high-ranked ones. So, for the price of one NFT, you’re actually getting a whole load of other NFTs that your AI will create for you. Awesome possum!
Below are three examples from our early tests as well as a video animation to show how the AiVATAR is “thinking” about the painting as it is created.
Above you can see three examples of the AiVATARs art creations. For the first the AI was asked to create art based around the theme of “Secret Garden”. The second piece was themed on “LGBTQ, Life and Planet”. The third was “Hive Mind”.
Finally, the video below shows the AI creating art involving a “sea monster”.
We have this planned for Phase 3 of the project, as you can see in the Road Map below. We have already built a pipeline for this, and it is indeed possible, but the user experience is really lacking because the tech and speed aren’t really there yet. Blame Gordon Moore for coming up with a shitty law 🙂 While it is indeed possible for you to ask questions and receive an answer, the current delay is too long and it doesn’t feel like a real conversation. For example, if you ask a question such as, “What am I doing wrong with my life?” – the AiVATAR will just sit there and stare at you for an uncomfortably long time before it answers. Technically, the bottleneck we are currently faced with (which will go away in time) is the hardware limitations that affect the speed of processing one AI and rendering another AI. All this will be solved in time though, we are confident.
For other cool ways you can use your AI, check the Road Map. It gets even cooler than just talking live to your AI… rent them out! Make copies to lease! Send them off to Zoom calls to discuss various things with other AIs and humans!
Q3 2021 - Launch AiVATAR Website and complete Minting. Verified previous Trey Collectors get 1 free AiVATAR in Pre-Launch. NOTE: In Phase 2, we have planned on your AI creating 10 pieces of custom AI based on the personality and word cloud of your AiVATARs. And yes, you can sell these as NFTs. It is possible we may add this to Phase 1 when we reach certain targets.
AiVATAR2 with many surprises. Hundreds of new AiVATARs based on actors we hire (the process is to get them to read a 15-minute script against a white wall and then we can have them say anything, in any accent). The AI will also have many different types of personalities instead of Phase 1 where they are all seeded with a "Philosopher" personality. Your AiVATAR will make you 10 pieces of AI art you can sell as NFTs based on their personality. NOTE: We already have this working and may add it to Phase 1.
Creation of AiVATAR V3 - Real time. Receive your own personal AiVATAR for real-time communication. Enable AiVATARS to communicate with one another. If your friend has an AiVATAR, just aim your phones at one another and enjoy the conversation. You could say, "Hey I'd you to talk to this AiVATAR about Trump," or maybe something a bit more lighthearted. 🙂 Enable owners of AiVATARs to send theirs off to live video Zoom meetings or Round-Tables where they can discuss topics with other AiVATARs and humans. Enable owners of AiVATARs to lease out copies of their AiVATARs for other people to enjoy. For example, say you got some attractive Scarlett Johansson avatar that is real perky and fun. You can let your friends borrow her for a while for a certain fee that you choose. Or, you can make 100 copies of her and lease them out at a daily (or nightly) rate.
When you buy an AiVATAR (you can buy between 1-20 randomly generated AiVATARs), they will be created by our server farm somewhere in Iceland and use a lot of geothermal energy to make a bunch of AiVATAR while people in Iceland are off having beer or rotten shark or whatever those people do when they are not shagging sheep. They don’t cost very much (the AiVATARs, not the sheep). It’s hard to say an exact dollar/pound/euro/etc equivalent value as ETH is all over the map but it will probably explode soon with ETH2. Let’s not get to crypto-esoteric.
You’ll also have to pay something called a “Gas Fee” which is yet another crypto-esoteric thing that I won’t get into here, but that fee can be anywhere from $40 to $150. Or $20 to $200. It’s hard to predict that part, but that’s all part of the fun too.
NEW: Ranking scores are now out. See the most rare ones with incredible answers here. You’ll notice some of the rarest ones are the AIs that tackle controversial topics and hit our rarity word cloud in a unique way! So cool.
Our rarity scale is based on many obscurities, including but not limited to: Trey’s photography in the background, the accent of the AiVATAR voice, human-based model we are calling a “Flesh Bag”, the position of the Flesh Bag, the size of the Flesh Bag, the length of the AiVATAR answer, voice type (e.g. Empathetic, Supa Zen, Like…your sexy mom, and about 20 others), and, finally, the word cloud of answers (most interesting). There will of course be unique combinations like the Santa avatar speaking in front of one of Trey’s photos of a snowy landscape. Santa is the rarest of Flesh Bags, so this has a very low probability when the AI generated the 6,000 AiVATARs. The calculations are incredibly complex and you can enter Our Discord channel if you want to know more.
We personally believe the most valuable AiVATARs may not be based on the traditional gamified rarity system where you have a Bored Ape that is wearing “Nihilistic Sunglasses,” but instead on how interesting the question/answer combination is. So, the value will be a bit more subjective, just like how humans choose friends.
Look, this is not a place to talk about your insecurities. Take those to your therapist. However, I can tell you how well-adjusted and emotionally stable people would answer this question.
I expect most collectors just to really love their AiVATARs as a totally unique creation the world has never seen. If you like one, then you may like to own more than one. It’s kind of like having children. You have two kids and, you’re like, well, that’s cool. And then you have a third kid, and you’re like errr there’s not really the economy of scale I predicted, and then you have a fourth kid and you’re like wow kids are super expensive and you really begin to wonder the upside/downside ratio of decisions you’ve made.
However, as opposed to the real world where it is somewhat difficult to sell your kids, you can actually sell your AiVATARs. So, imagine you had 10 kids and like 3 were really awesome and rare and weren’t gingers or always talking about veganism or something. Well, you can sell 7 of your kids make some sweet money ca$h. Note that this AiVATAR is not recommending any sort of human trafficking. It’s just an example.
In fact, Aivatars may indeed like a new owner. They might not have been that thrilled with you either, much like those 7 kids. Not that they have a say in it, as they are paraded around a big Ai auction but let’s not bring that up right now. Forget I mentioned it, my Aioverlord.
Frankly, even though many people will flip, we believe the value of these is for long-term collectors that know something special when they find it.
If you are still seeing the card back, that is very normal. This reveal is taking the longest in Opensea history because of the massive amount of data that we are uploading for your AI to be born. Most reveals happen in half a day, but this seems to be taking several days. There’s nothing wrong with the system, but our AI is melting the blockchain!
Well, here’s the blurb from the fine art website:
Trey Ratcliff is an artist on a somewhat quixotic mission to help spread consciousness and mindfulness to the world through photography and creativity. Running the #1 travel photography blog in the world, StuckInCustoms.com, has taken him to all seven continents over the past decade, and Google has tracked more than 140 billion views of his photos, all while building a social media presence with over 5 million followers.
Chris Anderson from TED called him a “pioneer” of the genre of High Dynamic Range photography. Ratcliff had the first HDR photograph to hang in the Smithsonian Museum and, sales of his large-format limited edition prints to fine art collectors worldwide have grown into a multimillion-dollar business. In 2012, he moved to Queenstown, New Zealand, before it became a trendy doomsday contingency plan. To see more quirky info, visit www.StuckInCustoms.com/Trey-Ratcliff
Trey has collectors of his fine art from all around the world. These include everyone from knights to professional football players to celebrities like Edward Norton and Leonardo DiCaprio.
View Trey’s media kit for additional information about the audience reach, partnerships, and other projects.
Yes! Anyone who holds one of Trey’s existing photography NFTs at noon UTC on Wednesday the 6th of October 2021 will be added to a whitelist which allows them early access to grab an AiVATAR for free (plus gas). You can then verify your ownership via Discord.
We recommend holding that NFT for at least a little while so that any delay on the whitelist generation doesn’t result in you missing out… really though we hope you like the image enough to keep it for a long time!
A number of those existing NFTs are sold out but available on the 2nd Market such as OpenSea. You can view them here and here, or on Foundation and some others on Makersplace. A wide range of pricing is available.
First, let us reiterate this important fact: Trey Ratcliff did NOT write any of the answers which are delivered by the AiVATARs. The responses are all 100% generated by artificial intelligence.
Trey created and curated the AiVATAR questions. There are about a thousand questions, so some are asked multiple times. We did this because we noticed the AI will often give different answers. Some of the answers are very interesting. Some of the answers are scary. Some of the answers are just plain weird. Some of the answers are hilarious. Some of the answers seem like exactly what a human would say. Some of the answers are very broken and it’s clear there is an AI speaking and that’s all part of the fun.
As with all the erratic writing on this website, Trey will alternate between the third and first-person as he finds his sense of “self” to be fluctuating based on his extended conversations with the AI.I’ll talk a little bit about the technology here, but if you want to know more, join our Discord channel and ask away!
So how does it all work? Well, in short we built a pipeline that is a beautiful mishmash of other technologies including GPT-3 from OpenAI and AI video content from Synthesia.
I (Trey) love the idea of creating art by using generative algorithms. My background is Computer Science and Math, but I’ve been a full-time artist for the last 15 years. So, the combination of the two is endlessly mesmerizing to me. How can we surprise and delight people and make them question reality? This is the artistic tension I enjoy exploring and using generative technology is, to me, mind-blowing.
The first step is coming up with the questions. This is a fun and daunting task. It’s much like the apocryphal three wishes to the genie. I can’t ask for material things from this genie, but I can ask deep questions, silly questions, weird questions, controversial questions, etc. And I don’t just have three wishes of questions, but infinite. I made a spreadsheet (exciting!) where I would create questions from scratch. I also found a ton of deep philosophical questions on the net, so I added and curated some of those. One of my hobbies is philosophy because I’m so curious about the nature of reality, mostly based on my psychedelic experiences. But that’s another topic. Anyway, I came up with about 1,000 questions and we would look at as many answers as we could. It turned out that some questions were great, and others produced pretty useless answers. I would delete those and add more good questions. There were a lot of iterations there, but I’m used to that as an artist. Try something. Fail. Try something else. Fail. Blah blah blah… Don’t beat yourself up if you fail. At least you tried. Get up, dust off, try again. Anyway, about 1,000 questions came out of this process.
Next, I would jam all these questions into a neural net AI and get a huge result back in text. These were stored in the cloud and then we would analyze the rarities such as accent, background, position, size, and other attributes. We created a word cloud of all the answers and I choose the words I find most interesting. For example, I would love to hear an AI talk about how to save the environment, so I would make “environment” a word and see how many AiVATARs mentioned it. Most of the words were articles like “a” “the” “and” etc, but those are not very interesting. So, I made the most interesting one of the rather cool elements of the rarity. However, the answer is still quite subjective, so, even if the AiVATAR mentions “Environment,” it might have said something like, “Sometimes a funeral is not the best environment for singing a show tune.” I personally believe that some of the most valuable AiVATARs will use a combination of words that I don’t expect, just like when you ask a friend a question and you get a completely surprising but delightful answer.
Next, now that we have all the text answers, we would jam them into a third-party video creator that can deep-fake real humans. What is a “deep-fake”? This is when you can use technology to make it appear that they actually said something. It’s an emerging technology, and it can be a bit scary. Note that none of the actual humans represented in these videos ever said these things. Many people have done deep fakes of Donald Trump, for example, where they make him say, “I really love Mexicans and I love Mexican food and I want as many Mexicans as possible to come into the United States.” When you watch a deep-fake video like that, it seems totally believable. There is a huge ethical question about all of this, but those are topics for deep, late-night jacuzzi philosophical chats. I won’t get into ethical issues here and I will try to take the high ground as the artist to create a venue, a tableau vivant if you will, for these sorts of discussions.
When we take the text and send it to the third party, it generates an MP4 file at 1080p, and this is the final file that will appear as the NFT. Each response is usually somewhere between one and two minutes. Sometimes more. It just depends on if the AI is feeling chatty. I can set a few parameters for how long the answer is. We typically go for a six “stanza” answer so the AI has a chance to wax-eloquent or completely go off the rails. One thing to note is that I have “embedded” all the questions inside of a larger context to seed the AI with certain ideas. For example, the FULL question might be “I am an AI that is very interested in philosophy, and I want to give you a considered answer for ‘[XXXX (What is the best way to share wealth around the world?)].” This of course brings up the possibility for AiVATAR2 where we can embed that question inside of another set of rare words, like a Madlib, and we’ll get some very different answers.
For this first version, I wanted to try to keep it kind, interesting, and provide deep answers. I’m sort of helping the AI self-identify as being “philosophical” so the answer would be deeper. And it worked. Well, in most cases. It’s clear that the AI has a mind of its own even if I try to seed it with an idea. It’s a lot like when you talk to a friend and assume their state of mind and say, “So, Harry, you are really an amazing athlete who is very healthy, and I was wondering [what is the best way to spend my day].” Your human friend Harry may say, “No, I’m not really much of an athlete, but I think it’s best if you start to learn more about the blockchain and get a huge portion of your assets into DeFi.”
It’s also worth noting that you’ll see a wide variation of backgrounds for these AIs. Those are all images from my photography portfolio and have an impact on the rarity of the NFT, just like the word cloud, etc. If one of those backgrounds appeals to you then check out my photography NFTs on OpenSea. Maybe you’ll find something you love.
Also, I should make it very clear, I am not a one-man team on this. I’ve got a great crew here and I am very grateful for the whole team. Anyone that knows me is completely aware that I don’t take myself seriously at all. Dissolution of the ego is one of the greatest things an artist can achieve. Going further, the team also includes our amazing community on Discord. We are listening to all of your feedback and are forever grateful. Remember, be kind. The AI is watching.
Trey Ratcliff is a photographer, artist, writer and adventurer as well as a Smithsonian featured fine artist with over 7million followers. His existing NFT collections have exceeded 500ETH in volume on OpenSea, FND and Makersplace.
His fine art is collected by people all around the world including from knights to professional sports stars to celebrities like Edward Norton and Leonardo DiCaprio. Trey’s images examine the tension between this reality and another. He is driven to help spread consciousness and love around the world.
Alex is passionate about changing the world in the best direction by using his deep knowledge of AI and the cloud to make cool shit. In other words, he’s a good-hearted supernerd. He has been obsessed with technology since before he was conceived, and this pre-coital passion has only grown stronger over time. His life goal is to use his extensive technical knowledge and experience to help all humans around the world. As with so many projects, he worries that the road to hell is paved with the best of intentions. This was written by Alex’s AI.
After university (Computer Science/Maths/Education) Stu moved into the finance industry and development of early internet banking/ecommerce platforms. A decade later he transitioned into technology journalism leading a publication with millions of monthly visitors and managing excellent working relationships with brands including Intel, AMD and NVIDIA.
Leaving his role as Editor in Chief, in 2016 he began working full time in Trey's creative empire. Fine Art sales, crypto management, render farm admin, music licensing, data wrangling, website design, social media admin and a sprinkle of After Effects/Premiere Pro work are all on his regular to-do list.
Tim Chang: Tim co-leads Mayfield’s Consumer investment practice and is an experienced investor and global executive. He has been twice named to the Forbes Midas list of Top Tech Investors.
Kevin Kelly: Amongst many things…Founding executive editor of Wired magazine and board member of the Long Now Foundation.
(Trey: Kevin is one my idols and he influenced me to start my blog with his 1,000 True Fans essay and his amazing TED talks, so I am beyond excited to have him on the board. Somebody pinch me.)
Hugh C. Howey: An American writer, known best for the science fiction series Silo. Published independently through Amazon.com’s Kindle Direct Publishing system. This early independent success allowed him to become a leader of the new paradigm of publishing and has since seen the series become a New York Times best seller. True to form Hugh again sits on the cutting edge of media, recently entering into agreement with Apple TV+ to create an adaptation of the Silo series.