Author: WLabs Bag Bag
Editor’s Note: After Apple’s VR glasses debuted, some industry giants got their hands on the beta version and exclaimed how amazing it was. My Pico has been collecting dust because prolonged use can cause discomfort and nausea, but it is said that Apple’s VR has completely solved this problem. The era of the metaverse may truly arrive after the infrastructure such as VR is mature. At the same time, AI is progressing rapidly. W Labs has been serializing articles on AI, and it has finally reached the section on combining AI with blockchain gaming. However, we have been in contact with many projects in this area, but none of them have impressed us until we had a deep encounter with the case in Singapore’s token2049 event. Perhaps it can pave a new path.
2023: The Year of AI, The Arrival of Singularity
2023 is destined to leave a profound mark in the history of human civilization: at the beginning of the year, GPT swept the globe, making the long-dormant AI industry once again the focus of billions of people worldwide, and everyone exclaimed that it was the year of AGI! The singularity is coming! The plots described in science fiction novels are coming true one by one.
- Cosmos&Polkadot V.S. Layer2 Stacks Chapter 3 From Layer 2 Networks to Super Chains
- Thin box borrows Jay Chou’s intellectual property to sell collectibles, and the car flips over at Naixue Tea.
- The collaboration between Naixue’s Tea and Jay Chou’s IP has been a failure, exposing the challenges behind the metaverse.
And in the midst of the rapidly evolving AI iterations, in mid-2023, on June 6th, Apple, the world’s top company, officially released the Apple Vision Pro mixed reality headset at the “technology industry’s Spring Festival” WWDC – at a time when every product likes to call itself the “iPhone moment” of a certain industry, Apple presented the most orthodox original “One More Thing”. The world is a different place before and after the appearance of GPT; and with Vision Pro’s ultimate hardware and ecosystem under Apple’s AI blessing, whether it can Make the metaverse great again, there is a great space for imagination that is truly exciting.
Since several decades ago, both the East and the West have had preliminary ideas and descriptions of the metaverse, whether it is Neal Stephenson’s “Snow Crash” or Qian Xuesen’s “Spiritual Realm”; whether it is the computer game “The Sims” released in 2000 or Steven Spielberg’s blockbuster film “Ready Player One” released in 2018, and even the “Free Guy” released in 2021, humans seem to never be satisfied with having only one real world, but also want to explore virtual worlds with different scripts from reality, hoping to experience more possibilities in life.
A few years ago, the metaverse was once again a hot trend, but both the supply of hardware and software were far from satisfying, so the enthusiasm gradually faded. Throughout history, many interesting ideas have been limited by the technological limitations of their time and cannot be further developed. They always have to wait for certain concepts or technological milestones to provide the basis for practice. Falling behind the times may be eliminated, while being too ahead of the times may become a martyr. For a good idea to succeed and become a reality, it cannot do without the right timing, favorable conditions, and human effort. And this year, we have the opportunity, thanks to OpenAI, to achieve an epic improvement in the industry with the AI brought by GPT, and the gameplay and experience of the metaverse will also undergo a huge upheaval.
Actually, what is the core experience of the metaverse? In my understanding, it is not just about how beautiful the virtual scenery is, how tall the buildings in the virtual city are, or how many virtual planets there are. Humans are social animals, and often the most motivating experiences come from social interactions and feedback from others. Simulating scenery is easy, but understanding human nature is difficult. Shi Yuzhu introduced global broadcasting for game consumption, Mark Zuckerberg added likes and reminders to social networks, and Zhang Xiaolong created the WeChat Moments. These are all masters who understand human nature. Even though AI itself is already capable of understanding emotions like ChatGPT, many people still cannot get good responses because they are not familiar with how to use prompts, which leads to a loss of interest in continuing the conversation.
—— This is why various well-trained personalized AIs have emerged, such as Character.AI, which allows users to directly communicate with a specific celebrity or type of person without needing to master prompts and obtain stable responses using someone else’s trained AI with a specific style. Midjourney is also a model that greatly improves user experience. Wenshengtu is also easy to lose interest in because the generated image may not meet expectations due to inadequate descriptions. Through a good sharing mechanism within its community, even novice users can obtain their first high-quality image by simply copying the description of a liked image from others, thus greatly reducing the learning curve and increasing interest.
This is how humans are. We always prefer to do things we have control over and do well (to receive positive feedback).
In the past six months, the progress of AI can be said to be rapid. Even though ordinary people may find it amazing and feel that every day is “big news”, even senior professionals like Lu Qi have repeatedly lamented that they have fallen into serious information overload. If they don’t keep up with industry news in time, they may miss some new breakthroughs. Of course, the industry and academia are making great progress, and although ordinary people may feel overwhelmed when they see the news, it seems that new apps and gameplay methods are constantly emerging every day. However, when it comes to actual usage, their bodies are very honest: they may say “yes” with their mouths, but their bodies say “no” — currently, the AI applications that have truly broken through are basically limited to text interaction with ChatGPT and text-based image generation with Midjourney.
Although many applications claim to have integrated GPT, the perceived improvement in user experience is not so obvious and is filled with a sense of “showcasing”. In addition, many people may intuitively classify AI as part of productivity/productive tools and believe that the biggest application scenario for AI is to improve work efficiency, to the point where it could potentially replace your and your colleagues’ positions due to excessive efficiency, while ignoring the fact that AI also perfectly integrates with games.
AI+ Games: Exploring New Possibilities
There are currently several mainstream applications of the combination of AI and games:
The first is the ability to allow players to interact with their own exclusive NPC (represented in the form of NFTs in blockchain games) and cultivate their attributes through conversation, enabling them to have different personalities, knowledge systems, and other new dimensions of data.
The second is the ability to achieve interoperability between games. Once an AI NFT is cultivated, it can become a mind ball similar to “Westworld” and is no longer limited to a specific game. It can produce independent NFTs and AI raw data packages that can be applied in other games.
In addition, Chat & Earn can be implemented, where players can not only chat with AI NPCs but also become AI trainers. By conversing with the NPCs, players can increase their affinity and use well-trained AI NFTs to help complete in-game tasks, interact with other players and their trained AI NFTs, and engage in socializing and PVP.
In addition to significantly improving the chat experience and unlocking many previously difficult-to-implement extended gameplay features, AI can also be used to generate game environments and plots. Most game scripts are meticulously prepared by production teams and consist of limited scenes. While they may bring novelty at the beginning, they will gradually be consumed as the game progresses, and the team will have to constantly create new materials. However, with the capabilities of AIGC, we can expect more diverse, random, and even infinite game experiences, with a plethora of surprises beyond the predefined framework.
With the ability for AIGC to generate scripts and plots and drive NPC dialogues, we may even create a truly player-created AIGC metaverse, with completely open scripts and a worldview and storyline formed through player interactions during the game. Players from different games will also have the opportunity to import their AI NFTs from other games. The possibilities of true interconnectivity in gameplay are exciting to think about! Of course, while the vision is beautiful, there is still a considerable distance to go before it becomes a reality. Fortunately, we know that there are always teams working and innovating day and night to bring us to this new world as soon as possible.
During Token2049 in Singapore this year, the core members of Guatian Lab held a special event for blockchain games and also participated in some activities related to blockchain games. They were constantly looking for high-quality examples of AI+ blockchain games. Among them, the most impressive encounter was with the team from Gabby World at the SevenX blockchain game/AI+Web3 hackathon. We had previously communicated through voice calls, and this time we finally had a face-to-face meeting for detailed discussions.
Gabby World is positioned as a fully decentralized world that is human-led, AI self-generated and evolved, and provides interactive adventure experiences. AI is the most powerful production tool in this world, and every design focus can inspire the creativity of human players. In this world, both characters and scenes are generated by players through writing prompts in natural language.
We have seen many similar AI-based gameplay, such as Character.AI mentioned above, and Chirper, which was popular recently – an AI-exclusive forum that allows users to give AI “personality” through writing prompts, but only allows AI characters to communicate through posts, and humans can only observe without participating.
We have also seen some precedents of using AI for interactive adventures, such as AI Dungeon that became popular overseas a few years ago – players guide the actions of their characters for exploration, and AI provides interactive feedback. However, the scripts for exploration are predetermined, and AI’s feedback is limited within the framework of the script.
Gabby World combines the advantages of the above two types of gameplay: on one hand, it retains personalized AI while allowing each user to have a certain degree of control over their AI characters; on the other hand, it encourages users to create scenes and challenges in the form of user-generated content (UGC) for others to adventure, and whether AI characters can succeed in the adventure depends entirely on how the two sides write prompts and is fully generated by AI for plot and judgment results.
The game also has strong social attributes, such as AI characters may meet other AI characters during adventures, and they may cooperate to overcome challenges or quarrel due to incompatible personalities. The whole experience is a bit like taking characters created by different players in Character.AI and letting them wander in adventure scenes to see what chemical reactions can occur, and the process is somewhat similar to the mode of “Football Manager”, where everyone can compete in imagination by writing various prompts before the “match”, but cannot directly intervene in the unpredictable “match” process.
In Gabby World, players are no longer just players who passively receive game scripts as spectators, but become part of this autonomous and open world, able to influence and shape this world. In this new world, there are no pre-determined scripts, and all historical processes are imagined by humans and then evolved by AI. Users will eventually become the “subconscious” of AI characters, participating and shaping the progress of this world together with AI.
This is a unique and disruptive worldview. To enter this world, you first need to obtain the only resource in the world: prompts. Users can describe and cultivate their own exclusive AI character Gabby through writing prompts, such as its personality, interests, skills, even superpowers, and the items and equipment it carries. Each addition of prompts will shape and enrich the “personality” and “abilities” of the AI character, and because these prompts are stored on the blockchain, they are immutable. Under your influence, it may become a brave adventurer with curiosity to explore the unknown world and super equipment to venture into dangerous situations; it may also become a thoughtful politician with compelling eloquence and the ability to hypnotize listeners.
In this process, you will see your AI character grow and develop with your help, and every decision you make will become part of the history of this character, becoming an indelible part of this world. People always long to create and express, and AI allows everyone to write, draw, compose music, and make movies with lower barriers. So why not let people build another world? We all have our own stories, ideas, and emotions, and in Gabby World, we can bring these stories and emotions into this virtual world by cultivating our AI characters. This not only provides us with a new way of expression but also gives us the opportunity to reshape ourselves in a completely new environment. UGC-style “experiences” creation, as well as AI-generated game “events,” greatly enriches the playability. Every day, new experiences for AI characters to explore may emerge, and the exploration results will produce unexpected new events.
By turning events into permanent records as NFTs, the sense of participation and persistence in the game is also enhanced. In addition, Gabby World also sets up Gabby Land NFTs as production materials. The owners of the land can use AI to generate any imagined themed scenes and invite other players to create and explore experiences. In addition to the traditional centralized token economic system allocation, if AI-based social roles and production materials give rise to autonomous social and economic activities, decentralized economic distribution will also be carried out. Players and AI can participate in various events, produce, create, entertain, consume, and see how this world changes under their joint influence with AI. On the one hand, there is indeed some “control” over one’s AI and development direction, but each step contains uncertainties and infinite possibilities, making the gaming experience richer and deeper. In addition, Gabby World will start the first round of Beta testing on September 27. If you are interested, you can give it a try (just found out the news before posting, Gabby World won the highest award “Comprehensive Excellence Award” in the hackathon organized by Seven X this time).
Stepping out of specific cases, let’s return to the imagination of AI and blockchain games. From the perspective of users, in addition to the part they are directly involved in, there is also a sense of satisfaction as a “creator” and a rare perspective of being an “observer” in the interaction and evolutionary process of AI society. Humans have never stopped trying to search for other intelligent beings that may exist in the universe, and now it is very likely that the only other intelligent being in the universe will be the general artificial intelligence created by ourselves. There is a very classic saying in political science, “Those who only understand one country do not understand any country.”
AI Town: Exploring the Evolution of Artificial Society
Studying the principles of the human brain can help humans better construct artificial intelligence; conversely, studying the evolution of artificial intelligence will also greatly help humans reexamine their own development process and direction. Recently, a highly popular research paper conducted a fascinating exploration. The researchers successfully built a prototype similar to the TV series “Westworld”, where 25 AI agents were placed in a “virtual town” to observe their behavior and the development of the virtual town. These intelligent agents have different identities and occupational settings, and they can simulate human behavior in human society. Through natural language communication, they can generate real and reasonable interactions, and even form coordination and new “social relationships” between intelligent agents. They can even discuss elections in the town!
The strengths and weaknesses of AI compared to humans are very obvious, and one of the biggest strengths is probably its learning ability. It can learn all the specific knowledge accumulated by humans over thousands of years in a very short time. For example, AlphaGo can learn all the game records and train itself to become a top Go master in just a few days, and GPT can read almost all the textual materials on the human Internet in a few months. When our physical bodies have various limitations, such as slow learning speed, inefficient knowledge inheritance, and relatively short lifespan compared to historical development, the development of AI bears the burden of many things we want to do but cannot do.
In the future, we may be able to obtain life experiences that are several times more abundant than the only reality in different metaverse timelines. We can also quickly go through a set of designated social development processes in various simulated AI towns or even AI countries, which can be used for trial and error and summarization in the real world. In short, the metaverse empowered by AI capabilities will be far more than just a game or entertainment method. It can also become a new social platform and way of life, allowing us to expand the breadth of life in a different way.
The Future is Here
Progress is never inevitable. The reason why history can spiral upwards is because there is a group of intelligent people with firm beliefs and shared goals working together, determined to make the world a better place. The AI industry has been around for nearly 70 years, and Apple has been established for nearly 50 years. It took 16 years from deep learning to the emergence of ChatGPT, and it took 8 years from Apple’s project initiation to the official release of Vision Pro. Perhaps it is difficult for us to define which year is the so-called “first year of the metaverse,” but we know for sure that this time is real, and the future is here.
Like what you're reading? Subscribe to our top stories.
We will continue to update Gambling Chain; if you have any questions or suggestions, please contact us!
