Today's Sector Market Cap: $13.04B +5%
The sector has rebounded significantly, with the crypto market warming up, leading to more volatile fluctuations.
📈24h Coingecko Trading Volume Top 10:
Token Trading Volume Market Cap Price Volatility
1- $VIRTUAL $159.2M $872.1M $1.33 +0.1%
2- $TAO $138.6M $3.7B $388.21 +6.6%
3- $AI16Z $56.6M $151.6M $0.14 +5.1%
4- $SHELL $55.0M $45.7M $0.17 -0.1%
5- $AIXBT $49.9M $128.9M $0.13 +5.6%
6- $SAHARA $43.9M $193.1M $0.09 -0.5%
7- $BANANAS31 $39.7M $72.0M $0.01 +0.1%
8- $KAITO $37.1M $290.2M $1.20 +4.3%
9- $GOAT $30.0M $106.7M $0.11 +9.5%
10- $WAI $28.7M $10.6M $0.04 -19.0%
⭐️ Key Information
1. Grayscale launched trust products targeting Walrus @WalrusProtocol
- Grayscale has launched two trust products for the Sui ecosystem: DeepBook Trust and Walrus Trust.
- DeepBook @DeepBookonSui is the leading DeFi liquidity layer on Sui, providing high-performance liquidity infrastructure;
- Walrus supports large-scale decentralized data management. @WalrusProtocol can achieve data storage, sharing, and real-time management on any blockchain, making it one of the important infrastructures for on-chain AI.
Comment: Sui is gradually establishing its influence among traditional financial institutions, and Walrus, as a leading solution in on-chain AI storage, is expected to gain more attention.
2. PrismaX @PrismaXai launched the remote training platform Teleop
- On the Teleop platform, users can remotely control robots, train visual-language action models, and earn PrismaX points.
- Points Rules: The first operation can earn 3,000 PrismaX points, and then every operation every 5 minutes can earn 90 points.
- Platform Link: https://app.prismax.ai/tele-op
Comment: Currently, the platform is relatively small, and training tasks may require queuing, but this remote training method with distributed user participation is highly innovative, both interesting and practical. Looking forward to future scaling and attracting more users.
3. Kunlun Wanwei @Skywork_ai released Matrix-Game 2.0 — the world's first open-source, real-time, long-sequence interactive world model.
- What is different about the world model: A 'virtual reality simulation' that the AI system can internally construct, which can learn to predict causal relationships and event developments in the environment, allowing AI to perform commonsense reasoning and decision-making like humans, rather than just relying on memory responses.
- Matrix-Game 2.0 uses the technology of 'one image generates ten thousand realms' to automatically generate videos of expanded 3D worlds from a single image.
- 25FPS: Single GPU achieves 25 frames per second generation. In videos, the higher the frame rate, the smoother the visuals. Previously, Google's closed-source world model Genie 3 had a generation speed of 24FPS.
- Minute-level interaction: Supports generating dynamic content continuously for several minutes, logically consistent, without interruption or delay.
- Application Scenarios: Game prototype development, AI agent training, virtual world generation, autonomous driving simulation environments, etc.
Comment: Domestic open-source models have surpassed top closed-source models in many aspects. The release of Matrix-Game 2.0 as the first open-source 'world model' has sparked enthusiastic discussions in the community.
📊 Data Source: @xhunt_ai, Cookie DAO, Coingecko
💡 Risk Warning: The above is for information sharing only and is not investment advice.
