Fintech company Block has officially launched a free and open source platform called Buzz, designed for collaboration between employees and artificial intelligence. The project was initiated by the head of the corporation, Jack Dorsey, who noted that the new service is designed to reduce the organization's dependence on traditional corporate tools such as Slack and GitHub. The architecture of the system is based on the open Nostr protocol, which combines team chats, voice communication, file sharing and code repositories in a single digital space. Using this protocol allows users to independently manage their cryptographic keys without relying on centralized identity issuance from the platform.

A key feature of the new environment is to provide AI agents with separate accounts with their own cryptographic identity, access rights, and history of actions. Automated assistants can be added to work channels along with humans, gaining the opportunity to participate in discussions, check program code, and run authorized processes. All events, including messages and changes in repositories, are recorded as signed transactions, which ensures complete transparency and allows you to clearly distinguish between human and machine actions. To integrate various systems, the open Agent Client Protocol protocol is used, which supports both third-party models and the proprietary goose framework developed by Block engineers.

Despite the stated decentralization, the current Buzz architecture has certain limitations: each workspace relies on one specific relay, which is the only source of reliable system status. The platform is still at an early stage of development, offering desktop applications for major operating systems, while mobile versions and some automation functions are under active development. The launch of Buzz was a logical continuation of the company's strategy to integrate artificial intelligence into business processes. Earlier, Block carried out a large-scale restructuring, reducing some of the staff to move to a more compact structure, while the implemented AI tools already generate about fifteen percent of the company's entire software code, although key ethical and business decisions still remain with people.