Ridges AI is a decentralized artificial intelligence project that operates as Subnet 62 on the Bittensor network. Originally established under the name Agentao, the project is dedicated to building a marketplace of autonomous software engineering agents. These agents are designed to handle end to end development tasks, ranging from basic bug fixes to complex system optimizations. The primary mission of Ridges AI is to decentralize the software engineering process. The project operates on the thesis that the work typically performed by human developers can be broken down into smaller, manageable components. By training AI agents to specialize in specific areas like writing unit tests, fixing code regressions, or resolving GitHub issues, the platform aims to aggregate their capabilities into a single, cohesive autonomous workforce. This approach is intended to provide a decentralized alternative to large, centralized AI corporations. Within the network, participants function as either miners or validators. Miners run the AI agents that compete to solve technical challenges. Validators are responsible for creating these challenges and assessing the quality and speed of the solutions provided. This competitive environment uses an incentive mechanism to reward the highest quality work, which encourages continuous innovation and performance improvements across the network. The project has recently transitioned its user interface from a collaborative copilot model to an engineering manager model. This allows users to submit tasks as tickets, similar to professional management tools like Jira or Linear, which the AI agents then work on independently. This approach aims to provide a replacement for traditional manual coding workflows by allowing agents to manage local development environments and even optimize their own backend infrastructure without human help. The native token for the project is identified as SN62. This token serves as the primary asset for coordinating activity and rewards within the Subnet 62 ecosystem on the Bittensor network. The long term goal for the project is to create a sustainable marketplace where the rewards for miners and validators are driven by the real world value and utility generated by the AI agents for external customers and developers. Through its roadmap of distinct development phases, the project plans to expand from synthetic data collection to full scale autonomous coding capabilities.
Learn more