The "Easy Button" For AI: Exploring FullJourney AI
Welcome to this week’s Deep-fried Dive with Fry Guy! In these long-form articles, Fry Guy conducts an in-depth analysis of a cutting-edge AI development or developer. Today, our dive is about Charles McCarthy’s project, FullJourney.ai, a complete suite of AI tools in one place. We hope you enjoy!
*Notice: We do not gain any monetary compensation from the people and projects we feature in the Sunday Deep-fried Dives with Fry Guy. We explore these projects and developers solely for the purpose of revealing to you interesting and cutting-edge AI projects, developers, and uses.*
🤯 MYSTERY LINK 🤯
(The mystery link can lead to ANYTHING AI related. Tools, memes, and more…)
Artificial intelligence (AI) has brought us image generation tools, video generation tools, voice generation tools, and more. Now it is time to bring them all under one roof, without the need to switch back and forth from tool to tool. This is the vision behind FullJourney.ai.
FullJourney is a suite of AI tools that encompasses all things media generation and editing, all in one place.
A FIGHTING START TO FULL-JOURNEY
Charles McCarthy, the co-founder of FullJourney.ai, has been dabbling in technology for the past 25 years. When he was a teenager, he worked for a digital marketing company. At that time, however, he was also training in mixed martial arts and landed a contract with the Ultimate Fighting Championship (UFC). He had a great career as a fighter and has since opened a few gyms along with a sports agency. After his retirement from the UFC, he came back to his roots—his love for computing, tinkering with technology, and building things.
For a while, he focused on virtual reality (VR), and was the first to build a live stream of a UFC event in VR. But as AI began to consume the technology scene over the past few years, McCarthy’s attention shifted. He said, “When I first started hearing AI come around again, it just felt like a buzzword at first—I didn’t think much of it—and then I tried ChatGPT and it blew me out of the water.” Through the emergence of ChatGPT, McCarthy notes, “AI went from being a cool, gimmicky thing about how you procedurally generate something to really feeling like you have a team that is working with you.” As a result of this experience, McCarthy’s curious mind began to wander, and he considered what he could build with this revolutionary technology.
As McCarthy explored different image generation platforms, he saw an underlying issue. He saw the potential of AI to generate images, speech, and even videos, but he wondered why some tools and applications could not do this all in one place. He thought, “If the images are so amazing, and ChatGPT can give me such great responses, why is all this other stuff not available in one place?” He began integrating some of his working experiments with large language models (LLMs), characters, image generation, jokes, and other ChatGPT integration on Discord. When he saw that people began playing with his experiments, he detected that he might be able to provide something substantive and useful for people.
McCarthy always took a particular interest in video generation. However, AI-generated videos have been bad for a long time—they have been slow, had poor resolution, and were quite unrealistic. When other people saw this as a negative, McCarthy embraced this as an opportunity. He thought, “How do I take this foundation and build on it to get something better out of it?”
This something better was manifested when McCarthy and his co-founder launched their text-to-video video platform with the development of Zeroscope v3, an open-source AI model. McCarthy wanted to push the project further and provide not only a text-to-video tool but an entire suite of tools. This led to developments of image and audio generation, speech, content editing, and more. McCarthy notes, “Any workflow that I could imagine would be a useful tool for media generation, I started incorporating into FullJourney.”
HOW DOES IT WORK?
As of now, the user can go onto FullJourney.ai and log in with their Discord ID. From there, the user can begin exploring the platform with a free monthly trial and explore different subscription options. The FullJourney website contains a list of all the relevant commands that can be used on this platform. New features have also become available where commands are less needed for media generation, and instead user has the option to create the content via chatbot conversations on the channel.
The process of leveraging the different abilities of the platform is quite seamless. For example, a user can ask the chatbot to make them an image of a frog wearing a hat. Once this image is generated, the user can prompt the chatbot to, say, “Change the color of the hat to red,” and the chatbot will do this. The user could then say, “Make this image into a video where the frog hops,” and the FullJourney chatbot will do just that. McCarthy notes something special about this process, which he enjoys using every day: “I love the randomness of AI. It takes what you want and mixes it with what it thinks you want, and this gives you something special each time.”
During this media generation process, users can drop into the help chat channel and ask for help if they need specific advice on maximizing certain prompts or on using the platform more generally. In the future, McCarthy plans on transitioning this suite of tools from Discord to a mobile app, making the experience even more streamlined from the consumer side, offering creative abilities at the user’s fingertips.
BEHIND THE PROMPTS
Each one of the AI media creations are done with different open-source AI models or models that have been fine-tuned or created by the FullJourney team with a few exceptions, such as the speech command which uses ElevenLabs application programming interface (API) and the DALL-E 3 command which uses the DALL-E 3 API. All of these models run on a coherent system which clusters models together so they can be used at scale. McCarthy notes, “In layman’s terms, we’ve got little AI guys that are in the cloud. You just tell them what you want.”
Most of the AI models in FullJourney utilize a process often referred to as diffusion, which works by throwing a lot of content together from images it was trained on and then uses word and pattern recognition to filter through undesired content until the user is left with what they want, given the context of their prompt. A more basic way to understand this process is that the models which use stable diffusion know that content with a certain caption usually looks like “fill-in-the-blank,” so that is what they produce when they are given a prompt.
In the near future, McCarthy is looking for ways to implement text-to-3D models on the platform as well as integrate conditional (if this, then that) commands, giving the users more versatility with their content creation.
THE FUTURE OF MEDIA GENERATION
Image generation tools have been becoming more and more advanced, but oftentimes these image generation tools must be used in combination with tools in other places to help the user get their desired content. FullJourney offers a one stop solution to this problem by offering video, image, and voice creation (along with more) on one platform.
FullJourney.ai is one of the most efficient media generation tools available, and as it continues to develop and gain traction, it will be used by people for a wide variety of purposes, including video creation, audiobooks, image generation, and more! The technology for revolutionary media generation is here, and the only limit is the human imagination.
BONUS: A LESSON FROM A UFC FIGHTER
Charles “Chainsaw” McCarthy, the co-founder of FullJourney.ai, has a unique perspective coming from a UFC background. He noted that this allows him to look at tech development from a different perspective than many others, which inspires him to build tools for the technologically savvy as well as for the layman. When he works with people in his gym, he envisions those people as his customers, saying, “I want any of my guys that I train with to be able to use it and create pictures of themselves climbing a mountain and put it on Christmas cards.”
One of the takeaways from his fighting experience that many can learn from is to take care of the body as well as the mind. This helps McCarthy in his development and gives him a fresh perspective and energy when he faces challenges in computing, and he is an advocate for others in the tech space to do the same.
McCarthy’s experience in fighting has also given him confidence to overcome obstacles he faces in programming. He is used to being in tricky positions and in spots where he feels like the odds are stacked against him, and in fighting and in programming, his mindset is one of persistence and overcoming obstacles to reach his goals. When he faces roadblocks, he remarks, “Give up is not an option. I’m not going to quit. I’m going to get it done; sometimes it is just a matter of rethinking my approach until I smash through that barrier. Everything is a temporary limitation, and everything has never been done until it’s done.”