Weekly Challenge

WIN META AI GLASSES!! CHALLENGE 3 - SECOND NATURE

Build AI that understands the moment. Second Nature is a hackathon focused on building context-aware AI products—software that understands what a person is doing, recognizes what is happening around them, and provides useful assistance at the right moment. Most software requires users to communicate through clicks, forms, and carefully written prompts. This hackathon challenges participants to create more intuitive products: AI that can interpret context, adapt to changing situations, and help users accomplish meaningful tasks with less friction. Your challenge is to build a working AI-powered product that uses context to deliver a more relevant, adaptive, and useful experience. Context may come from:     •    Images, video, or a device’s camera     •    Voice, conversation, or environmental audio     •    Documents, files, and on-screen content     •    A user’s current task, activity, or workflow     •    Information shared across applications and tools     •    Preferences, history, or changing user needs The strongest projects will not simply wait for a perfect prompt. They will understand enough about the user’s situation to determine when, where, and how to help. When We Say “Context-Aware AI,” We Mean - Build AI that can perceive - Create products that understand visual, audio, textual, or application-level information—not only what a user types into a chat box. - Build AI that understands intent - Use context to recognize what the user is trying to accomplish, even when every detail has not been explicitly provided. - Build AI that adapts - Design an experience that changes its guidance, interface, or behavior based on the user, task, and situation. - Build AI that takes the next step Go beyond generating an answer. Build software that can guide a user through a task, coordinate tools, identify something important, or take a useful action with appropriate user control. Start with a clearly defined user and problem. Context should make the product meaningfully more useful—not merely more technically impressive. Participants may build:     •    Web or mobile applications     •    Desktop products     •    AI agents and workflow tools     •    Voice or conversational experiences     •    Vision-powered applications     •    Accessibility and assistive products     •    Developer tools or infrastructure     •    Creative and educational experiences     •    Interactive software prototypes Projects must be software-based and should run using capabilities available on standard consumer devices. No custom hardware or electronics are required. 🤖 Use of AI AI must be central to the product’s core functionality or user experience. A generic chatbot, basic prompt wrapper, or isolated AI-generated feature will not be sufficient on its own. The project should demonstrate how AI uses context to enable an experience that would be difficult, impractical, or impossible with traditional software. Every submission must include:     •    A working prototype demonstrating the project’s core experience     •    A demo video of no more than 90 seconds     •    A clear real-world use case identifying the intended user and problem     •    Feedback from at least one test user     •    At least one improvement made—or planned—based on that feedback     •    A brief explanation of what context the product understands and how it uses it     •    A short note addressing relevant privacy, safety, accessibility, or reliability considerations Judging Criteria Usefulness — 25% Does the product solve a meaningful problem for a clearly defined user? Context-Aware Experience — 25% Does the product use context in a way that is relevant, thoughtful, and essential to the experience? Originality — 20% Does the project introduce a fresh product idea, interaction model, or application of AI? Execution — 20% Does the prototype work, and is the core experience clear, reliable, and well demonstrated? User Learning and Responsibility — 10% Did the team respond meaningfully to user feedback and consider privacy, accessibility, safety, and potential AI failures? What We Are Not Looking For     •    A generic chatbot with a different system prompt     •    A thin interface around a single API call     •    A familiar product with AI added only for novelty     •    A dashboard that only displays generated content     •    A concept presentation without a working core experience     •    Context collected without a clear benefit to the user Responsible Use Teams should collect and use only the context necessary for their product. Projects that process cameras, microphones, personal files, activity, or other sensitive information should clearly communicate what is being accessed and why. Users should remain informed and in control of important actions.