Here are specific examples of subnets in the Bittensor (TAO) network and what they do — that is, the tasks their AI models and applications solve:
🔹 Examples of active subnets and their functions
1. 📚 SN1 — Apex (language models)
Created for training and using large language models (LLM) — answers questions, translates texts, assists with explanations, programming, and other tasks similar to the work of ChatGPT.
2. 🧠 SN2 — Omron (verification through ZK proofs)
Evaluation and verification of AI outputs using zero-knowledge proofs, which helps ensure the correctness and reliability of responses.
3. ⚡ SN4 — Targon (speedy responses)
Optimizes outputs of language models — fast and accurate text synthesis for applications requiring prompt response generation.
4. 📊 SN10 — Sturdy (DeFi optimization)
Focuses on autonomous algorithms for managing yield in DeFi (decentralized finance).
5. 🗄️ SN13 — Dataverse (distributed data)
Stimulates the collection and storage of distributed data — useful for building databases and analytics.
6. 🌐 SN18 — Cortex.t (generation and API access)
Provides decentralized access to models generating texts and images via API interfaces.
7. 🖼️ SN23 — SocialTensor (content creation)
Generation of text, images, memes — that is, content for social applications and creative tasks.
8. 🧠 SN32 — It’s AI (AI content recognition)
Identifies which data was generated by AI and which was not — helps in the fight against fake content.
9. 🔍 SN34 — BitMind (deepfake detection)
AI subnet for finding and classifying deepfake videos and fake content, helping to increase trust in digital information.
10. 🔢 SN43 — Graphite (graph problem solving)
Solves labor-intensive graph problems (for example, optimal path problems), useful for scientific, logistical, and engineering applications.
11. 🧑💻 SN45 — Gen42 (code generation)
Specializes in automatic code generation, which helps developers solve programming tasks faster.
🚀 Other interesting areas
The network operates with 128+ subnets, each with diverse goals.
AI models operating directly on devices (federated learning, confidential data).
Synthetic generation of identities for testing security systems.
Rapid indexing and real-time data feeding for AI agents (for example, Subnet 82 — Hermes).
📌 Briefly about the role of subnets in TAO
Each subnet is a separate market for artificial intelligence within Bittensor:
It has its own reward rules and methods for evaluating model outputs.
Miners produce AI outputs, Validators evaluate them, and the network distributes TAO rewards for successful results.
$TAO #TAO/USDT
