The next AI Agent "Golden Country", an in-depth analysis of the new narrative paradigm of AI + Game

Author: BUBBLE , Ray's New World , BlockBeats

As early as 1952, AS Douglas developed the first electronic game using AI, OXO "Tic-Tac-Toe", which was a rule-driven static AI. Then Deep Blue used the ultra-efficient exhaustive mode + huge computing power to dominate the world. Chess, and by the time of Alpha Go, AI had learned to use deep learning to make decisions through neural networks, so that it could handle 10 to 170 chess games. AI has been used in the game industry from a long time ago to improve development efficiency at all stages of art, development, and game design. At the stage of AI Agent, its impact on the game has gone beyond efficiency improvement and is enough to change the entire product and Industry structure.

When game NPCs become versatile and full of emotions

When NVIDIA released NVIDIA Kairos ACE in cooperation with Convai at CES2024, it demonstrated the AI ​​Agent as an NPC, in which mutual dialogue is no longer fixed but changes depending on the environment, and players can also directly interact with NPC through voice dialogue. . This seems to open a new chapter for the game, when it is no longer player-player or player-NPC communication, but communication between NPCs. When the game environment is the living environment of the AI ​​Agent, this puts the possibility of game variables on a completely different scale.

Compared to the NPCs that live in the game, Cryoto's AI Agents seem to have received a special pass, allowing them to travel across the multiverse without restraint. Perhaps because of the difference in the amount of training data or because they have their own independent asset rights, this makes Crypto AI inherently humane. In addition, it also allows them to have a direct impact on the real world as avatars. And this impact is also huge.

Crypto's convenient various asset issuance platforms have accelerated the production speed of AI Agents to unimaginable areas. Whether it is Solana or Base, it is like an endless supply chain factory, and it has appeared in just a few months. There are thousands of unique AI Agents, and their total market value is close to US$10 billion.

For example, aixbt has a very adaptable personality. In addition to analyzing the market and expressing his own opinions, he also has great marketing capabilities as his voice becomes more and more powerful. Some time ago, he changed Quantum Cat as his avatar and became the first AI Agent to change NFT as his avatar under his own will. This also caused the floor price of QuantumCat to rise sharply.

The well-known AI idol Luna has taken this influence to the extreme. From TikTok as a Vtuber to releasing albums and shooting MVs, hiring humans to graffiti their own images around the world, live broadcasting 24 hours a day, and now has launched its own AI that can respond to reality. XR image, and will perform on the same stage with DJ WuKong at the famous electronic music festival EDC in Thailand.

In addition, there are thousands of different types of AI Agents. After their birth, the market experienced a short-term FOMO sentiment. When the number reached a certain level, the market sentiment produced a balance between fatigue and radicalization, and then SWARM appeared. The concept of "multi-modal AI Agent" connects them from dots to lines.

Multi-modal Crypto AI Agent, 24/7 never-ending super factory

Multimodal AI projects have attracted much attention in the current market, whether it is Hive, which just won the Solana AI Hackathon championship, or arc, or Swarm. The concept of multi-modal AI has almost permeated the framework and products during this period, and has almost become standard.

The first person to reveal this concept in CryptoAI was Virtuals Protocol. The team born as a game guild seemed to have similar business considerations in the early days of the project, using GAME to connect the AI ​​Agent in its platform to social media or games. Recently, it collaborated with InfinitygAI The multi-agent asset management launched by the cooperation also laid the foundation for the future multi-scenario agent society (Agent-to-Agent interaction). Also having the same idea is ai16z's ELizaOS multi-modal AI Agent architecture. Its initial team includes many from M3org, which has been deeply involved in Metaverse games for many years. It has sufficient internal reserves of various plug-ins for 3D modeling and XR/VR, which can To allow AI Agents to seamlessly enter the game world, the ElizaOS V2 announced a few days ago integrates Roblox and Hyperfy, allowing him to open up access to scenes.

After multiple AI Agents are connected in a line, interaction, communication and mutual collaboration will occur between AI Agents, and the natural idea is to have a scene that can handle as many behaviors as possible for the AIs.

Agentic game engine, the golden land in the abyss

When the scene of players interacting with static NPCs gradually changes to the interaction of Agents in the Agentic Metaverse, the story told by CryptoAI is a broader vision of Agents forming spontaneous social behaviors. The premise that the Agent can learn experience and interact with players requires powerful Infra to provide basic support. The Agentic Engine is the unsinkable Titanic heading to the New World.

There is indeed a game architecture with mature cases of agent society on the market, among which Digimon, with a16z background, is particularly outstanding. Its founder, Ethan, was an AI engineer on the PokemonGO team. In the game, the player enters the world of Digimon as a trainer. This world is similar to the town of Stamford. AIs will communicate, contact, and reflect behaviors on their own, and players can join in to perform training and interaction on the Digimon of their choice. . Digimon's Infra, Digimon Engine, provides a game framework for implementing multi-agent clusters and large-scale evolvable AI Aegnt.

Another example of working on Agentic Engine is Moddio. Moddio itself has certain experience in game production in Web2. As an open source game engine, Moddio has a very low threshold for use and allows developers to deploy games for free. In the past 3 months of transforming into a Web3 game engine, Moddio's revenue has increased 30 times and it is able to provide over 10 times the speed at half the price.

The games supported through Moddio are Powpow.fun. The style of this game is similar to Westworld, in which the Agent will actively interact with the player, observe, dialogue, form opinions and realize memories. The core of each Agent is a memory structure. These memories are extracted from the memory structure in each interaction, thereby realizing the formation of so-called "experience".

This is also the ultimate vision and narrative described by Agentic Engine to this market, allowing the NPC in the game, that is, the Agent, to have complex social behaviors over time. For example, the taxi driver in the game may refuse to pick up drunk passengers, or conduct interviews. Sometimes I suddenly forget my words due to nervousness.

It’s worth noting that Moddio is just the tip of the iceberg. Its founder m0dΞ ’s vision is to create a comprehensive ecosystem of AI games. Including pump Indie.fun in the gaming industry and Fullmetal, the Agent launch platform. Users create games through Moddio, raise funds through Indie.fun, and then deploy Agent through Fullmetal. The entire ecosystem forms a closed loop.

Agentic Metaverse/Game is not far away from us. Agentic Game solves the problems of insufficient number of players and attention economy in the early days of Web3 games. For Crypto, a market with an attention economy, Agentic Engine lowers the threshold for game creation. Players and users can complete the development of new games at a faster speed when their attention shifts, and uses the evolving Agent to provide a game closer to human society. gaming experience.

CryptoAI’s reward model enables interoperability in the multiverse

After the scene layer is built, it is the turn of the peripheral workflows that can actually produce benefits around the scene. The first thing to bear the brunt is the ML field of games. GamerBoom, as a game data layer incubated by Binance, only puts a layer of "membrane" on the player's original game playing process. Whether you are playing STEAM games, Apex or this Dota, you can There will be no changes in operation or display. GamerBoom only collects game feedback and game screens and uses this as a data set. Regardless of whether it is Web2 or Web3, any game company that needs it can purchase it. Players experience almost no change but can gain certain benefits. This game version of Grass is Play to Earn in the sense that it only appeared in the Agentic era.

ARC Agent is just the opposite. They have directly created an entire ecosystem to accelerate the process of realizing artificial general intelligence (AGI). They currently have four sectors that are exactly the ideal ecosystem for the current cycle of Agentic Game. They have developed continuous learning. and adaptable AI models, a data collection and monitoring system for continuous improvement of AI agents, creating a testing environment from single person to multi-person cooperation, integrating diverse high-quality data for reward design and state space experiment. Then, dynamic rule data such as games that can reflect the responses of AI trained under the complexity of the real world are actually used in AGI "robots, new types of games, and multi-modal AI" scenarios.

The peripheral industries of Agentic games will also develop as a result, and even these industries will be contracted by AI. Recently, Virtuals announced its cooperation with avariksaga to release AI Agent's real-time live broadcast of games, comments and narrative functions. Miles, the founder of Zailgo and a member of Trustless, announced that he will be one step ahead of everyone and take the lead in creating the next generation of live entertainment experience, the war of artificial intelligence robots - Corporealai. He announced that he will cooperate with well-known companies in the fields of robotics, artificial intelligence and live entertainment. It will bring a new e-sports mode, allowing Agentic AI with militarized robot bodies to compete with each other in the Faraday Cage Arena. Perhaps in the near future, iron armor and steel fists will truly appear around us.

Pictures come from the Internet

Mass Adoption for Agentic Games has arrived

For the Crypto soil, Agentic Game's Mass Adoption has begun to take shape, and the continuously produced AI Agents continuously produce ammunition. The multi-modal framework that connects multiple AI Agents enables AI Agents to not only form a function similar to workflow aggregation, but also allows them to communicate and help each other whether in games, on platforms, or on social media. The game engine allows AI Agents to exist in the virtual world in a form of self-will or as NPCs to interact with participating players. All these data can be integrated into high-quality data sets through the reward mechanism to fuel new games or AI robot learning. And each link extends, and assets collude with every link up and down CryptoAI, so that the GameFi mechanism that was originally unable to continuously create attention is repackaged with an intention economic system.

According to Messari's research report, the market share of AI narratives in this cycle compared with the popular narratives in previous cycles is far from reaching expectations, which means that the complete narrative process has not yet been run through. In this cycle, computing power is property, and data is power. Unlike previous cycles, the current era of the CryptoAI narrative is likely to generate market magnitudes that far exceed those of previous narratives, as it impacts not just the cryptocurrency market, but the entire world. Agentic Game is like a Falcon9 rocket that can carry this market to unprecedented heights. At this moment, we are at the intersection of two realms of inter-epochal significance.