Biteye AI Daily September 15

Today's zone market capitalization: $22.22B

The weekend is calm and peaceful, the decentralized computing model continues to be validated, and the focus of development in the robotics field is shifting from hardware to software. Decentralized data collection may usher in a new round of explosion. Top 10 24h Coingecko Trading Volume: Token Trading Volume Market Capitalization Price Volatility 1- $ATH $329.0M $791.0M $0.07 +26.1% {future}(ATHUSDT)

2- $ai16z $276.1M $129.5M $0.12 -9.8% {future}(AI16ZUSDT)

3- $OPEN $262.0M $211.2M $0.98 +0.1% {spot}(OPENUSDT)

4- $VIRTUAL $106.8M $831.2M $1.27 -5.6% 5- $TAO $64.6M $3.4B $351.20 -1.3% 6- $IP $56.7M $3.0B $9.63 -0.8% 7- $PROMPT $29.4M $40.2M $0.18 -4.6% 8- $SAHARA $26.9M $202.1M $0.09 -1.4% 9- $GOAT $25.4M $105.6M $0.11 -5.1% 10- $KAITO $24.6M $285.4M $1.18 -2.6%

⭐Key Information

1. Vader @Vader_AI_ launched the smart glasses mission platform EgoPlay.

• EgoPlay is a gamified platform where users can earn Vader Points by completing daily tasks using smart glasses.

• The collected data can be directly used for robot training, and the current daily tasks include activities such as folding clothes, washing dishes, making sushi, playing golf, pushing wheelchairs, operating a drill, and playing the piano.

• The platform supports various brands of smart glasses. As long as the video meets the requirements, it can be uploaded, and videos can also be directly uploaded from the phone's photo album.

• Currently, EgoPlay is still in the closed testing phase and participation is limited to whitelist users. Whitelist eligibility is for users who have previously participated in staking and currently still hold tokens or have staked on Virtuals.

Comment: The development focus of the robot track is shifting from hardware to "intelligence," with data collection becoming central to training intelligence. EgoPlay adopts a more inclusive approach that makes it easy for ordinary users to participate, which is worth continuous attention.

2. Sportstensor @sportstensor announced a strategic partnership with @Polymarket.

• Sportstensor is Bittensor's subnet number 41, focusing on sports predictions. It allows miners to deploy AI models to predict match outcomes and enhances accuracy by weighting the best predictions through aggregating meta-models. Over $2 million in rewards have been allocated to contributors so far.

• The specific cooperation direction has not been announced yet, but at this stage, the combination of AI and prediction markets is mainly reflected in: AI-assisted decision-making for betting, AI Agent for automated trading and market making, as well as AI generating and expanding new prediction markets, etc.

Comment: Against the backdrop of the current booming prediction market, although the details of the collaboration have not been disclosed, the price of Sportstensor tokens rose by about 40% after the announcement of the partnership, reflecting the market's high expectations for the integration of AI and prediction markets.

3. Alibaba Qianwen has released a new open-source model Qwen3-Next series.

• Includes two versions:

Qwen3-Next-80B-A3B-Instruct: Designed specifically for instruction optimization, suitable for dialogue and task-oriented applications. Qwen3-Next-80B-A3B-Thinking: Supports 'Thinking Mode', suitable for tasks that require reasoning or step-by-step analysis.

• High inference computation efficiency: Using MoE structure, the model has a total of 80B parameters, activating only about 3B parameters, significantly improving computation efficiency.

• The processing capability for long text tasks has significantly improved: natively supports a context length of 262K tokens, expandable to 1M tokens.

• Compared to Qwen3-32B-Base, the training cost is only 10% of it, and the inference throughput has increased by 10 times.

Comment: The reduction in training costs and the improvement in inference efficiency of Qwen3-Next make it more practical for ordinary small and medium developers, lowering the development threshold and expanding application scenarios.

Data source: @xhunt_ai, Cookie DAO, Coingecko

Risk warning: The above is for information sharing only and is not investment advice.