What exactly is AIGC that everyone is talking about?

Author: One0x, CryptoFinders

What is AIGC?

AIGC (AI Generated Content) refers to the use of artificial intelligence to generate the content you need, where GC stands for creative content. Correspondingly, there are well-known concepts such as PGC, which is content created by professionals; UGC, which is content created by users themselves; and AIGC, which is content created using artificial intelligence.

For example, the currently popular ChatGPT is a relatively successful product within the AIGC domain. Once launched, it swept the globe, surpassing 100 million monthly active users in just one month, a figure that is certainly more attractive than the historical data of popular products like TikTok and Facebook.

In addition to the Q&A field represented by ChatGPT, AIGC has also launched many excellent products in other areas. One of the more mature developments is in the AI painting field, where AI creates images based on your instructions.

Currently, the AIGC sector has become one of the hottest new trends, with major companies making strategic moves. If the metaverse in 2021 was referred to as a "bubble," then the rise of AIGC will drive the development of the metaverse and help it become a reality.

In the entire digital ecosystem, if the metaverse is seen as the representative of Web 3.0, then virtual application scenarios are the most important output content of the metaverse. Celebrities like Elon Musk and Mark Zuckerberg are also involved, investing heavily to create new virtual applications.

When AIGC is integrated into the metaverse, it will undoubtedly become an important part of content production. Of course, AIGC has broad and diverse functions and will not be limited to a single industry. Currently, the first thing that comes to my mind is that its visual aspect will be the first to produce results.

Development History of AIGC

Tracing back to 2014, the proposal of "Generative Adversarial Networks" (GAN) became a hot deep learning model among major companies that year. In retrospect, this can also be considered the earliest practical framework for AIGC.

The basic principle of GAN adopts the idea of a game between two entities, where the generative model G continuously generates and outputs data, which is then input into the discriminative model D along with the training set for judgment, followed by optimization learning.

The generative model G and the discriminative model D engage in a game, learning from each other to achieve optimal results (the output of the generative model, when input into the discriminative model along with the training set, results in an output of 0.5, meaning it cannot determine whether the input is real data).

The above image shows the process by which the generative model learns to transform a given input X from a uniform distribution to a normal distribution.

Fast forward to 2020, Web3+AI successfully emerged, with the generative art NFT leader Art Blocks being a successful application case. It represents the first step of AIGC in the blockchain field.

Art Blocks is a platform for randomly generating artworks, founded by Erick Snowfro, focusing on programmable, generative content. The content it generates is immutable on the Ethereum blockchain.

Its random process is managed by a string of numbers stored in NFTs on the Ethereum network, which control a series of attributes of the artwork you purchase, ultimately generating a unique NFT according to your ideas.

For creators, they need to adjust and deploy their generative art scripts in advance on Art Blocks, ensuring that the output matches the input, and then store the scripts on the ETH network using tools.

For collectors, when they mint a series of works, they essentially receive a random hash value, and then the script executes, creating an NFT corresponding to that hash value.

Just before this article was published, CZ announced on Twitter that Binance's first AI product, Bicasso, has officially launched. Users can use AI technology to generate new NFTs from their personal works after adding relevant descriptions. This is also the first application of "image-for-image" in the blockchain field.

Before 2021, AIGC primarily generated text (ghostwriting), while the new generation of models can handle various formats, including text, sound, images, video, actions, and more. It can fully leverage technological advantages in creativity, expressiveness, iteration, dissemination, and personalization.

In 2022, the development speed of AIGC was astonishing. At the beginning of the year, it was still in a rudimentary stage, but just a few months later, it reached a professional level, with product results that could easily be mistaken for real.

Future Outlook of AIGC

From the proposal of the concept to the maturity of the product, AIGC took ten years.

The maturity of AIGC also makes the realization of the metaverse no longer an empty talk. It can genuinely help the future development of the metaverse while significantly reducing labor costs. By leveraging AI to break the shackles of the production process, the limitless application and imagination allow the metaverse to develop rapidly, producing high-quality content efficiently.

We should believe that given the current state of development, it is possible that one day in the future, a virtual chip will suddenly be launched, enabling users to explore infinitely in the virtual world. At that time, we may find ourselves, much like today with ChatGPT, embracing curiosity and a novel experience as we truly step into the era of Web 3.0, ushering in a new era of the metaverse.