How AIGC realizes commercialization — NFT case

In 2022, AIGC is considered to be a new type of content creation after PGC and UGC, and this year, the corresponding technology and applications are piling up, among which the deep learning model continues to improve, the promotion of open source model, large model to explore the possibility of commercialization, become the “acceleration” of AIGC development. 2022 has expected to catch the momentum of the first year of AIGC(AI Generated Content).

Mr. Hu Bo, a popular artist on Maze AI’s Chinese platform, was originally a well-known short film director, having directed the famous Xinjiang intangible cultural heritage documentary and cultural tourism promotional film. He has also tried his hand at creating cyberpunk future hyper-realism and quadratic style, which has attracted a lot of fans.Its release of NFT has the number one commercial conversion rate across our platform.

In the following, we will introduce the journey of the artist Mr. Hu Bo to the AIGC art track and the experience of the commercialization model of AIGC from the perspective of an interviewer. For more interaction, please leave your ideas in the comment.

1. When did you first encounter the concept of AIGC? Can you tell us about your understanding of it?

The first time I saw the American game designer Jason Allen’s “Space Opera” drawn with midjourney won the first prize at the Colorado State Fair art competition, and I also started to think about finding some tools for drawing 3D graffiti to expand my creating path. This August, I swiped Maze AI, a generation tool on Shake, and found that it could directly generate images with some fields. Later, I participated in Maze AI’s whitelist activities one after another, applied for creators, and went from research, creation, and participation to NFT series sales. This is roughly my AIGC experience.

In addition, AIGC has also lent me a great hand in my daily work. The inspiration of art itself comes from the accumulation of visuals, and AI as a tool to give full play to imagination, some of the generated artwork drawings have a great inspirational effect on me. For example, when I do some fashion styling films, I will use AI tools to render a sketch first, and often these sketches are more effective than my imagination, which can be borrowed when we shoot including light effects, makeup, etc.

2. You have been using Maze AI since you first started, and now you are generating fine quality diagrams, what problems did you encounter and how did you solve them?

Yeah, the first AIGC tool I used was Maze AI China version. Later I also tried other AIGC tools, some too tech some slow generation speed, some out of the picture effect is not ideal, subsequently gave up using. The Maze AI China version is relatively easy to use, only a paragraph of descriptors to generate pictures, get started quickly, the generation speed is generally less than 10 seconds, and the results are also more ideal. In addition, Maze AI China supports both English and Chinese languages, which is also convenient for Chinese people.

I encountered a lot of problems in this process. At first, I was just thinking like a normal person, inputting ordinary written statements, similar to the format of subject-predicate object, like writing a middle school essay, which was a bit difficult, and the final result was not very good. Later I realized that AI has its own logical weighting for reading, and the pattern of statements suitable for it is usually “phrase + comma separator + phrase”. The first step is to transform the human-readable statements into the AI-readable ones.Once the transition is over the follow-up will be very good to get started and progress will be fast.

For example, if I need to create a landscape, then I need a description of the picture, from a distant view, a medium view, to the subject, then retouching, including the season, the weather to describe this picture, while listing the desired artistic style, color aspects, you can also draw on the style of some artists, to create, this way the generated work quality will be closer to what you want.

3.The “Star Kindergarten” series NFT works you released are of high quality and selling relatively well, and many users take your works as their avatars, how do you ensure that the works are of high quality and the style has a high relevance?

The “Star Kindergarten” series is my attempt to do NFT through AIGC, and I didn’t expect it to be so successful. The market for NFT actually started last year, but there is still a threshold for ordinary people, and it just so happens that AIGC has lowered the threshold for ordinary people and can be quickly out of the picture, the times are changing, the creation of the natural way to push the new.

4.Do you have any advice for ordinary creators on the direction of commercialization of AIGC?

In terms of my own practice, the first thing is to find a field you like, the style you want to generate.Then being hard to train your own generation techniques, you can go to accumulate keywordsas well, but if you want to develop your own style, it is still inseparable from a lot of practice.
Another significant thing is to insist on outputting high-quality AI content with your own style. Of course art workers can’t detach themselves from the original intention of making art creations. When your starting point and content touch some compatible market perception, you will be a big sale.

Over the past few decades, the potential of AI art has increased significantly, soaring all the way to the top with results everywhere.Since the release of Maze AI in its first generation, we have been working hard to optimize the models. We have already included anime, SD DD, Avatar models, etc. In the future we will have open models.We also sincerely hope more artists will come in to stimulate the creative potential of both sides and gain win-win benefits.