Coinspeaker
Insights into Web3 Gaming: How AI Creates the New Genshin Impact

“Genshin Impact” has been a game changer in the video game industry. Seemingly coming out of nowhere and released to the world in September 2020, developer miHoYo has created an open-world, multi-platform game which has become one of the highest-grossing mobile games of all time, and the highest-ever grossing video game for its first year of release. Many developers aspiring to emulate the success of “Genshin Impact” may end up scratching their heads, asking the question, how on earth can its success be replicated?

Potential investors may be asking the same question. Looking back to their humble surroundings, the only external investment during miHoYo’s first year of operation in 2012 was an angel investment of $150,000, while today it is worth an estimated $7 billion. But where is the next miHoYo?

Can the open world of “Genshin Impact” be replicated in Web3? The GalaxyBlitz team believes that with the winning combination of SLG+AI, the success of “Genshin Impact” can be replicated in the age of web3.

Let’s first take a look at the payment system of “Genshin Impact”. Players’ sentiment for the in-game characters is the main reason players spend money. It is ultimately the emotional bond that players develop for their characters why players are willing to part with their money in the game.

miHoYo has created highly three-dimensional and complex characters with which players can experience and explore the game together with, and gradually come to know and understand, just as you would a real person. It is for this reason that players develop these bonds with the characters and want to go on this emotional journey through the game together with them, that players are so willing to spend their money in the game.

This is also the reason why “Genshin Impact” is regularly updated with new content. Only regular high-quality content updates centered around the main characters can keep player’s interest and cause them to develop empathy and intimacy with the characters in the game, thus encouraging them to regularly spend money on in-game items.

SLG in Web3 Will Create a New “Genshin Impact” through AI Technology

So, can such a payment system be replicated in web3 strategy games? Yes, through the development of AI technology. Once AI technology is applied in the GalaxyBlitz game, all soldiers in the game will become AI characters with whom the player can resonate with. All the actions of the soldiers, such as dodging, defending, walking, sprinting, bending, and shooting are all dictated by the gamer. Through the application of AI technology, each AI character will behave differently from one another when in battle.

By cooperating with the renowned AI company, Phantom AI, GalaxyBlitz has successfully developed the GalaxyBlitz Mobile Net, a neural network training system suitable for strategy mobile games. It is a scalable, data-driven method for AI learning under the constraints of low power consumption by mobile devices.

An overview of the framework:

GalaxyBlitz Mobile Net

The GalaxyBlitz Mobile Net framework consists of two parts: pre-training and task training.

During pre-training period: the low-level policy π(a|s, z) is defined as a mapping of the target “z” in the action space dataset “s” with “a” as the base action as a variable, and the reward function “r” is used for various actions. After pre-training, it can be transferred to the task training by using a specific task.

The low-level policy can model various general actions, and the training of general actions such as sword swinging, walking, sprinting, bending over, and shooting can be realized according to the player’s habit status set defined by Environment.

During target training: high-level policy w(z|s, g) is defined as a map that achieves target g using action combination z as a variable in action space dataset s, where z = array(a).

As shown in the figure below, through the high-level policy, the AI ​​Being who has learned the low-level policy can be instructed to perform various specific tasks. These specific tasks are connected by a set of actions in the low-level policy.

Action task combination of high-level policy – kneeling shooting bullseye

In the high-level policy, at the beginning, the simulated character performance skill combination generated by the random sample of Environment1, with the frequent operation of the player, the discriminator will separate out the actions that the player is used to, and the actions of AI Being will become more and more in line with the player’s habit.

Our actions do not include “lying on the ground” and “standing up”, but AI Beings will learn “lying on the ground” and “standing up” autonomously. If the player likes the timid character of AI Being, then AI Being will “fall down” in the face of danger and “stand up bravely” at critical moments, so as to discover the shining point in the player’s heart.

The Rise of the New SLG for Web3.0

If the current version of the game achieves an annual turnover of 200 million US dollars, its economic performance will reach similar levels as high-performing SLGs in the web2.0 field. But as AI will be implemented in the next GalaxyBlitz version, the player’s investment, entertainment, and emotional consumption needs will be met.  Therefore, it is anticipated that GalaxyBlitz can eventually reach an annual turnover of $5 billion similar to the level of “Genshin Impact”.

GameFi’s evolution path as envisioned by the GalaxyBlitz team

Powered by GalaxyBlitz’s AI technology, this next generation SLG will have thousands of playable characters as NFTs. In the game, each character will embody the player’s unique preferences and actions. They are no longer merely playable tools as in traditional games. Instead, the character’s emotions, playability, and scalability can all grow and evolve as the game progresses.

Therefore, the AI ​​Being full of personality will bring the narrative climax of the next Web3 game.

Insights into Web3 Gaming: How AI Creates the New Genshin Impact