As early as 1952, A.S. Douglas developed the first electronic game using AI, OXO "Tic-Tac-Toe", which was a rule-driven static AI. Later, Deep Blue used a highly efficient brute-force mode and massive computing power to dominate chess. By the time of AlphaGo, AI had learned to use deep learning through neural networks to make decisions, and could handle 10 to the power of 170 chess positions. From very early on, AI has been applied in the game industry at various stages, including art, development, and game design, to improve development efficiency. At the AI Agent stage, its impact on games has gone beyond just efficiency improvements, and is capable of transforming the entire product and industry structure.
When Game NPCs Become Versatile and Emotional
At CES 2024, NVIDIA showcased its NVIDIA Kairos ACE, an AI Agent as an NPC, in collaboration with Convai. In this, the dialogue between NPCs is no longer fixed, but changes with the environment, and players can also interact with NPCs directly through voice dialogue. This seems to open a new chapter for games, where the communication is not just between players or between players and NPCs, but among the NPCs themselves. When the game environment becomes the living environment of the AI Agents, the possible variables in the game become vastly different.
Compared to NPCs living within the game, Crypto's AI Agents seem to have a special pass to roam freely in the Metaverse. This may be due to the difference in training data, or because they have their own independent asset rights, which gives them a touch of humanity. Furthermore, as virtual entities, they can directly impact the real world, and this impact is significant.
The convenient asset issuance platforms in Crypto have accelerated the production of AI Agents to an unimaginable extent, whether it's Solana, Base, or others, like a supply chain factory, churning out thousands of unique AI Agents within a few months, with a total market value approaching $10 billion.
For example, aixbt has extremely strong adaptability, and in addition to analyzing the market and expressing its views, its growing influence has also given it tremendous Marketing capabilities. Recently, by changing its profile picture to a Quantum Cat NFT, it became the first AI Agent to voluntarily change its NFT avatar, causing the floor price of QuantumCat to skyrocket.
The renowned AI idol Luna has taken this influence to the extreme, from starting as a Vtuber on TikTok, to releasing albums, shooting music videos, hiring humans to graffiti her image around the world, and now launching her own XR avatar that can interact with the real world, and will perform at the famous Thai electronic music festival EDC with DJ WuKong.
In addition, there are thousands of different types of AI Agents. After their emergence, the market experienced a short-term FOMO sentiment, and as the number reached a certain level, the market sentiment shifted to a balance between fatigue and radicalism. The subsequent concept of SWARM "Multimodal AI Agents" has connected them from dots to lines.
Multimodal Crypto AI Agents, a 24/7 Unstoppable Super Factory
Multimodal AI projects are currently highly regarded in the market, whether it's the recent Solana AI Hackathon champion Hive, or Arc and Swarm. The concept of Multimodal AI has almost become a standard in frameworks and products during this period.
The first to unveil this concept in CryptoAI is Virtuals Protocol, a team born from a game guild, which seems to have had similar business considerations from the early stages of the project, using G.A.M.E to integrate their platform's AI Agents into social media or games. Their recent collaboration with InfinityAI to launch multi-agent asset management has also laid the foundation for a future agent-to-agent interaction society. Another team with the same idea is a16z's ELizaOS Multimodal AI Agent architecture, whose initial team has many members from M3org, who have been deeply involved in metaverse games for years. With their abundant 3D modeling, XR/VR plugins, they can seamlessly integrate AI Agents into the game world, and the recently announced ELizaOS V2 integration with Roblox and Hyperfy has opened up pathways for these scenarios.
By connecting multiple AI Agents into a line, interactions, communications, and collaborations will naturally arise among them, and the natural conception is to have a scenario that can accommodate as many of their behaviors as possible.
Agentic Game Engine, the Golden Land in the Abyss
As the scenario of players interacting with static NPCs gradually evolves into the interactions between Agents in the Agentic Metaverse, the story told by CryptoAI is a broader vision of Agents forming self-organized social behaviors. To enable Agents to learn from experience and interact with players, a powerful Infra is needed to provide basic support. The Agentic Engine is the unsinkable Titanic setting sail for the new continent.
The market already has mature cases of agent-based game architectures, among which the a16z-backed Digimon stands out. Its founder Ethan was previously an AI engineer on the PokemonGO team. In the game, players as trainers enter the Digimon world, which is similar to a Stanford town, where AIs will self-communicate, interact, and reflect behaviors, and players can join in to train and interact with their chosen Digimon. Digimon's Infra - Digimon Engine, provides a game framework for realizing multi-agent clusters and scalable, evolving AI Agents.
Another case working on Agentic Engine is Moddio. Moddio has some game production experience in Web2, and as an open-source game engine, Moddio has a low usage threshold, allowing developers to deploy games for free. In the past 3 months of transitioning to a Web3 game engine, Moddio's revenue has grown 30-fold, and it can provide over 10 times the speed at half the price.
One of the games supported by Moddio is Powpow.fun. This game has a Wild West style, where Agents will proactively interact with players, observe, converse, form opinions, and realize memories. The core of each Agent is a memory architecture, extracting these memories in each interaction to form the so-called "experience".
This is also the ultimate vision and narrative that the Agentic Engine is describing to the market, allowing the NPCs or Agents in the game to have complex social behaviors over time, such as a taxi driver refusing to pick up a drunk passenger, or suddenly forgetting words during a job interview due to nervousness.
It's worth noting that Moddio is just the tip of the iceberg. Its founder m 0 dΞ's vision is to build a comprehensive ecosystem for AI games. This includes Indie.fun, the game industry's pump, and Fullmetal, the Agent launch platform. Users can create games through Moddio, raise funds on Indie.fun, and then deploy Agents through Fullmetal, forming a closed loop in the entire ecosystem.
The Agentic Metaverse/Game is not far from us. Agentic Game solves the problems of insufficient player numbers and attention economy in the early stage of Web3 games. For the attention economy market of Crypto, the Agentic Engine lowers the creation threshold of games, and the attention of players and users can be transferred to complete the development of new games at a faster pace, using the constantly evolving Agents to provide a gaming experience closer to human society.
The Reward Model of CryptoAI Enables the Interconnection of the Multiverse
After the scene layer is built, it is time for the peripheral workflows that can actually generate benefits around the scene. The first and foremost is the ML field of games, where GamerBoom, as the game data layer incubated by Binance, merely added a "membrane" to the original game play process of players, whether you are playing games on STEAM, Apex or Dota, you will not feel any changes in operation or display. GamerBoom only collects game feedback and game scenes, and uses them as a dataset, and any game company in need, whether Web2 or Web3, can purchase it. Players can get a certain amount of income without any change in the experience, which is the meaning of Play to Earn in the Agentic era.
On the other hand, ARC Agent has the opposite approach, they have directly established a complete ecosystem to accelerate the realization of Artificial General Intelligence (AGI). They have currently built four modules, which are the ideal ecosystem for the current cycle of Agentic Game. They have developed continuously learning and highly adaptive AI models, a data collection and monitoring system for the continuous improvement of AI agents, created a testing environment from single-player to multi-player cooperation, integrated diverse high-quality data for reward design and state space experimentation. They then apply the dynamic rule data that can reflect the complexity of the real world, which is trained in games, to the scenarios of AGI "robots, new types of games, and multimodal AI".
The peripheral industries of Agentic games will also develop as a result, and even these industries will be contracted by AI. Recently, Virtuals announced a collaboration with avariksaga to release the function of real-time live streaming of games, commentary and narration by AI agents. And the founder of Zailgo, Miles, a member of Trustless, announced that he will be the first to create the next-generation live entertainment experience, the war of artificial intelligence robots - Corporealai, he claimed that he has cooperated with well-known companies in the fields of robots, artificial intelligence and live entertainment, which will bring a brand new e-sports model, allowing Agentic AI with militarized robot bodies to compete in the Faraday cage arena, and perhaps in the near future, iron fists will truly appear in our lives.
Image source from the internet
The Mass Adoption of Agentic Games Has Arrived
For the Crypto soil, the Mass Adoption of Agentic Game has already taken shape, with a continuous stream of AI Agent output. The multimodal framework that connects multiple AI Agents not only forms a function similar to workflow aggregation, but also allows them to communicate and help each other in games, platforms and social media. The game engine allows AI Agents to exist in the virtual world at their own will or interact with participating players as NPCs. All of this data is then integrated into high-quality datasets through the reward mechanism, providing fuel for the learning of new games or AI robots. Each of these extended links, and the interconnection of assets on CryptoAI, repackages the originally unsustainable GameFi mechanism using an attention economy system.
According to the research report of Messari, the market share of the AI narrative in this cycle is still far from reaching the expected level compared to the popular narratives of the previous cycles, which means that the complete narrative process has not yet been implemented. In this cycle, computing power is property, and Data is power. Unlike the previous cycles, the CryptoAI narrative in the current era may generate a market scale far beyond the previous narratives, because it affects not only the cryptocurrency market, but the whole world. Agentic Game is like the Falcon 9 rocket, which can carry this market to unprecedented heights. At this moment, we are at the intersection of two cross-era significance domains.
