Ocean

**MCP (Model Context Protocol) is establishing itself as the standard interface layer for agentic AI systems.** Key developments: - **Unified access**: Once an agent connects to an MCP endpoint, it eliminates the need for provider-specific APIs - **Streamlined workflow**: Agents can now discover compute resources, check real-time pricing, execute jobs, and retrieve results through a single interface - **Agent-native design**: This represents a shift toward infrastructure built specifically for autonomous AI operations The protocol enables agents to function as modular components within the AI stack, allowing any LLM to discover and execute functions without custom integration work. This standardization marks a move away from treating AI agents as simple chatbots toward recognizing them as independent operational units. The development suggests a maturing infrastructure layer for AI systems, where interoperability and standardized protocols replace fragmented, provider-specific implementations.

The Singapore Court of Appeal has confirmed that confidentiality protections no longer apply to emergency arbitration proceedings involving Fetch. The court agreed with the lower court's assessment that **Fetch's own conduct** made it unreasonable to expect confidentiality would remain intact. **Key developments:** - Court of Appeal upheld the High Court's original decision - Interim confidentiality orders have been discharged - Emergency arbitrator's award is now publicly available The full arbitration award document has been [republished here](https://drive.google.com/file/d/1G7rb3z6VUq82u2eQsikkMrqhD3FTkKtE/view?usp=sharing). This ruling sets a precedent for how confidentiality can be waived in arbitration proceedings when a party's actions undermine the expectation of privacy.

**Ocean Network launches premium NVIDIA H200 GPU access** Ocean Network has made NVIDIA H200 GPUs available through their dashboard at $2.16 per hour. The service includes: - **Pay-per-use pricing** with escrow-protected payments - **Free compute credits** for testing - **Remote execution** on global nodes - **Direct IDE integration** via Ocean Orchestrator The platform addresses common cloud computing pain points by charging only for actual execution time. If a node fails, no charges apply. If code fails, users pay only for compute actually used. Key features include: - Pre-qualified, benchmarked nodes - Containerized job execution - Real-time monitoring and logs - Fast recovery when nodes go offline Users can launch AI workloads directly from VS Code, Cursor, Antigravity, or Windsurf editors without managing infrastructure or dealing with idle billing. [Try Ocean Network Dashboard](https://dashboard.oncompute.ai/) [Read Documentation](https://docs.oncompute.ai/ocean-orchestrator/using-ocean-orchestrator-with-ocean-dashboard)

**Lunor AI** has launched the **TripFit Tags** data annotation challenge, running until March 10 with a prize pool of **1,500 USDC**. **How it works:** - Read travel listings and review their details - Categorize each listing by traveler type: Solo, Couple, Family, or Group - Help train smarter travel search systems The challenge offers straightforward tasks that contribute to improving travel recommendation algorithms. Participants label data that will enhance how travel platforms match listings to user preferences. [Learn more about the challenge](https://twitter.com/lunor_ai)

Ocean Protocol has launched Ocean Nodes, a decentralized computing infrastructure designed for AI and machine learning workloads. **Key features:** - Builders can access geographically distributed compute to train, fine-tune, and run models without centralized cloud providers - Ocean C2D (Compute-to-Data) keeps data and algorithms sealed inside containers—compute executes remotely, only outputs are returned - GPU owners can monetize idle hardware by contributing capacity to the network - Jobs run in isolated, containerized environments The infrastructure aims to create a sovereign compute layer for AI that is open and distributed, addressing the growing demand for GPU resources as AI workloads scale. [Learn more about Ocean Nodes](https://docs.oceanprotocol.com/developers/ocean-node)

Ocean Protocol has introduced a **Free Compute feature** to their VS Code extension to address image generation failures during rapid scaling. The new feature allows users to: - Start with small, fixed-seed baselines - Lock their settings for consistency - Save both configurations and outputs - Maintain stable, repeatable runs This approach helps developers **experiment and scale at their own pace** while avoiding common pitfalls that occur when scaling too quickly. Users receive **7,200 seconds of free compute time** to test the feature and get started with their projects. The extension continues Ocean's mission to provide seamless AI development tools that combine compute power, privacy, and algorithms directly within developers' preferred IDE environment.

**CivicLens annotation challenge completed** with 206 contributors analyzing European Parliament speeches for political discourse patterns. **Key achievements:** - Labeled speeches for stance, claims, tone, topic, and ideology - Partnership with @lunor_ai delivered high-quality political data - Results provide insights into political discourse analysis The challenge focused on creating datasets to help AI models detect market-moving political statements more effectively. [Read full details](https://blog.oceanprotocol.com/annotators-hub-civiclens-turning-speeches-into-signals-e27041c3972c)

The **CivicLens challenge** ends today at 16:00 UTC with a **$10,000 prize pool** still up for grabs. Participants help build datasets for AI model training that enables compliance teams to: - Detect checkable statements faster - Identify market-moving information more accurately This data annotation challenge focuses on improving AI systems used in financial compliance and regulatory oversight. [Join the challenge](https://app.lunor.ai/challenge/1000039) before time runs out.

**Ocean Protocol has officially released its VS Code Extension** on the Visual Studio Code marketplace, making decentralized compute accessible directly from developers' coding environment. **Key features include:** - Run compute jobs and train AI models without leaving your IDE - Direct access to Ocean Nodes for decentralized computing - Built-in real-time monitoring for distributed AI workflows - Privacy-preserving Compute-to-Data job deployment The extension eliminates the need for separate dashboards, integrating **compute and data access directly into developers' workflows**. This streamlines the process of building federated learning systems and distributed AI applications. Developers can now tap into Ocean's decentralized infrastructure without switching between multiple tools or interfaces. The extension supports safe deployment of privacy-preserving AI models while maintaining the familiar VS Code development experience. [Try the Ocean VS Code Extension](https://docs.oceanprotocol.com/developers/vscode)

Ocean Protocol introduces Data NFTs, enabling researchers to tokenize and manage scientific intellectual property on-chain. Key features: - Mint sequences, models & datasets as verifiable assets - License genes for research with datatokens - Offer API access to models - Split royalties via smart contracts - Track versions & access logs The system provides automated IP management through smart contracts, replacing traditional legal processes. All assets are timestamped, auditable, and wallet-linked for seamless compliance and publication tracking. Learn more at [Ocean Protocol Docs](https://docs.oceanprotocol.com/developers/contracts/data-nfts)