Original author: TechFlow
The crypto market was devastated after this week’s “Black Monday,” but tokens in different sectors rebounded a day later.
Among them, the most popular one is Bittensor (TAO).
Coinmarketcap data shows that among the top 100 tokens by market value yesterday, Bittensor (TAO) rose 23.08%, ranking first on the rebound list.
Although the AI narrative is not as hot as at the beginning of the year, the choice of hot money also represents optimism about the leading projects in the sector.
However, Bittensor has also suffered a certain degree of fud before. The community believes that the project is overrated and there is no practical application in the subnet.
Although the usefulness of a crypto project is not directly related to the token price, is Bittensor really just an empty shell?
In the past few months, 12 new subnets have been added to Bittensor, and each subnet has promoted AI-related development to a certain extent, and new Alpha projects may emerge among them.
We took a look at these new subnets to see how their fundamentals are changing while all attention is focused on TAO’s price rebound.
Subnet 38: Sylliba, a text-to-speech translation tool that supports 70+ languages
Development Team: Agent Artificial
Introduction:
Sylliba is a translation application that supports both text and voice translation and can handle more than 70 languages.
It is worth mentioning that this program can be used by on-chain AI agents:
Automated translation process: AI agents can automatically call this service to achieve cross-language information processing and communication.
Enhance AI capabilities: Enable AI systems that do not have multilingual capabilities to handle multilingual tasks.
Translation requests and results can be verified on the blockchain, adding credibility to the system.
Incentive mechanism: Through the token economy, high-quality translation service providers can be incentivized.
Project address: https://github.com/agent-artificial/sylliba-subnet
Subnet 34: Bitmind, detecting and distinguishing real content from fake synthetic content
Development team: @BitMindAI
Introduction:
BitMind is focused on developing decentralized deepfake detection technology. As generative AI models rapidly advance, distinguishing high-quality synthetic media from real content is becoming increasingly complex.
BitMind’s Subnet solves this problem by deploying a powerful detection mechanism in the Bittensor network, using both generative and discriminative AI models to effectively identify deepfakes.
Meanwhile, the BitMind API enables the development of powerful consumer applications that leverage the deepfake detection capabilities of the subnet. A BitMind web application with an image upload interface can use the API to help users quickly identify the likelihood that a picture is real or fake, providing an easily accessible and easily interpretable anti-spoofing tool.
Subnet 43: Graphite, intelligent path planning network
Development team: @GraphiteSubnet
Introduction:
Graphite is a subnet specifically designed to work with graph problems, with a particular focus on the Traveling Salesman Problem (TSP). The TSP is a classic optimization problem where the goal is to find the shortest possible route that visits a set of cities and returns to the starting point.
Graphite leverages Bittensor’s decentralized machine learning network to efficiently connect miners to handle the computational needs of TSP and similar graph problems.
Currently, validators generate synthetic requests and send them to miners in the network. Miners are responsible for solving the TSP using an algorithm of their design and sending the results back to the validators for evaluation.
Subnet 42: Gen 42, GitHub’s open source AI coding assistant
Development team: @RizzoValidator, @FrankRizz 07
Introduction:
Gen 42 provides decentralized code generation as a service leveraging the Bittensor network. Their focus is on creating powerful, scalable tools for code-based question answering and code completion driven by open source large language models.
Main products:
a. Chat Application: Provides a chat front-end that allows users to interact with their subnets. The main functionality of this application is code-based Q&A.
b. Code completion: Provide an OpenAI-compatible API that can be used with continue.dev.
For details on how miners and validators can participate, please refer to the project Github
Subnet 41: Sportstensor, sports prediction model
Development team: @sportstensor
Introduction:
Sportstensor is a project dedicated to developing decentralized sports prediction algorithms, powered by the Bittensor Network.
The project provides basic models on the open source HuggingFace for miners to train and improve, while enabling strategic planning and performance analysis based on historical and real-time data, and rewards comprehensive data set collection and high-performance predictive model development.
Miner and validator functions:
Miners: Receive requests from validators, access relevant data, and use machine learning models to make predictions.
Validator: Collects miners’ predictions, compares them with actual results, and records verification results.
Subnet 29: coldint, niche AI model training
Developer: Not found yet, the official website is here
Introduction:
SN 29 coldint, the full name is Collective Distributed Incentivized Training.
Goal: Focus on pre-training of niche models. Niche models may refer to models that are not as widely used as large general models, but are very valuable in specific fields or tasks.
Participation and division of labor among miners and other roles:
a) Miners are primarily incentivized by publicly sharing training models.
b) Secondary incentives are given to miners or other contributors who share insights by contributing to the code base.
c) By rewarding small improvements, miners are encouraged to share their improved work regularly.
d) Highly reward code contributions that combine individual training efforts into better combined models.
Subnet 40: Chunking, optimizing the dataset for RAG (Retrieval-Augmented Generation) applications
Development team: @vectorchatai
Token: $CHAT
Introduction:
SN 40 Chunking is like a very smart librarian. It divides a large amount of information (text, pictures, sounds, etc.) into small chunks. This is done to make it easier for the AI to understand and use the information. If the bookshelf is well organized, you can find it quickly.
SN 40 Chunking is helping AI organize its bookshelves.
Not only text, SN 40 Chunking can also handle multiple types of information such as pictures, sounds, etc. It is like an all-round librarian who manages not only books, but also photo collections, music CDs, etc.
Subnet 39: EdgeMaxxing, optimizing AI models to run on consumer devices
Development team: @WOMBO
Introduction: S N3 9 EdgeMaxxing is a subnet focused on optimizing AI models for consumer devices, from smartphones to laptops.
The EdgeMaxxing subnet uses a competitive reward system with daily competitions to encourage participants to continuously optimize the performance of AI models on consumer devices.
Participant roles and division of labor:
Miners:
The main task is to submit optimized AI model checkpoints
They use various algorithms and tools to improve model performance
Validators:
Must run on specified target hardware (e.g. NVIDIA GeForce RTX 4090), collect models submitted by all miners every day, benchmark each submitted model against a baseline checkpoint; score based on speed improvement, accuracy maintenance, and overall efficiency improvement, and select the best performing model of the day as the winner
Project open source repository: https://github.com/womboai/edge-maxxing
Subnet 30: Bettensor, a decentralized sports prediction market
Development team: @Bettensor
Introduction:
Bettensor allows sports fans to predict the outcomes of sports games, creating a decentralized sports prediction market based on blockchain.
Participant roles:
Miner: responsible for generating prediction results
Validator: Verify the accuracy of the prediction results
Data Collector: Collect sports event data from various sources
Project open source repository: https://github.com/Bettensor/bettensor (seems to be still under development)
Subnet 06: Infinite Games, a general purpose prediction market
Development team: @Playinfgames
Introduction:
Infinite Games develops real-time and predictive tools for prediction markets. The project also arbitrages and aggregates events from platforms such as @Polymarket and @azuroprotocol.
Incentive system:
Using $TAO tokens as incentives
Rewarding providers of accurate predictions and valuable information
Overall, the project encourages users to participate in forecasting and information provision, forming an active forecasting community.
Subnet 37: LLM Fine-tuning, Large Language Model Fine-tuning
Development team: Taoverse @MacrocosmosAI
Introduction:
This is a subnet focused on fine-tuning Large Language Models (LLMs): miners are rewarded for fine-tuning LLMs and model evaluation is performed using a continuous stream of synthetic data from subnet 18.
Working mechanism:
Miners train models and publish them to the Hugging Face platform regularly.
Validators download models from Hugging Face and continuously evaluate them using synthetic data.
The evaluation results are recorded on the wandb platform.
TAO tokens are rewarded to miners and validators based on weight distribution.
Project repository address: https://github.com/macrocosm-os/finetuning
Subnet 21: Any to Any, creating advanced AI multimodal models
Development team: @omegalabsai
Introduction:
Any to Any in this project refers to the ability of a multimodal AI system to convert and understand between different types of data or information, such as text to image, image to text, audio to video, video to text.
The system can not only perform conversions, but also understand the relationship between different modalities. For example, it can understand the connection between a text description and an image, or the connection between a video and its corresponding audio.
In this subnet, incentive mechanisms are used to encourage AI researchers and developers around the world to participate in the project. Specifically:
Contributors can earn token rewards by providing valuable models, data, or computing resources.
This direct financial incentive makes high-quality AI research and development a sustainable endeavor.
Project repository address: https://github.com/omegalabsinc/omegalabs-anytoany-bittensor
Additional knowledge:
In case some readers don’t know the significance of Bittensor subnet, a simple explanation can be:
A subnet is a specialized network in the Bittensor ecosystem.
Each subnetwork focuses on a specific AI or machine learning task.
Subnets allow developers to create and deploy AI models for specific purposes.
They use cryptoeconomics to incentivize participants to provide computing resources and improve models.
Original link
