Block just laid off more than 4,000 workers. Now the fintech company wants to change how teams talk to each other. And build software. And let AI agents sit right beside their human colleagues in the same digital room.
The effort comes in the form of Buzz. A new platform unveiled this week that Block positions as an alternative to Slack, GitHub and a handful of other workplace tools. It runs on the decentralized Nostr protocol. It gives every participant – person or bot – a cryptographic identity. The whole thing is free, open source and self-hostable. Block thinks this setup delivers better auditability, data control and security than proprietary chat apps.
“Every company is going to need a place where humans and agents work together,” said Bradley Axen, head of AI capabilities at Block, in the company’s official announcement. “The question is whether that place is proprietary or open. We built Buzz because we believe it should be open.”
The timing raises eyebrows. Block’s staff cuts, announced earlier this year, slashed its workforce by roughly 40 percent. From more than 10,000 employees down to under 6,000. CEO Jack Dorsey tied the reductions directly to advances in artificial intelligence. A smaller team armed with better tools could achieve more, he argued at the time. Shares jumped on the news.
Yet here Block stands, pushing forward with an ambitious new product built around those same AI capabilities. The company spent the past two years developing and releasing open-source AI tools. Those experiments led to one clear conclusion. The most effective work doesn’t come from asking an AI for help in isolation. It happens when humans and agents share context. When they operate in the same space.
Buzz tries to create exactly that space. Channels. Threads. Direct messages. Voice calls. Media sharing. Code repositories. Automated workflows. The interface will feel familiar to anyone who has used a modern team communication tool. But the agents in Buzz aren’t simple chatbots waiting for commands. They hold their own cryptographic key pairs. They receive defined permissions. They post updates, review code, run approved automations and join conversations as full participants.
Multiple agents and multiple humans can share one workspace. They build on each other’s contributions rather than operating in silos. Block’s GitHub repository for the project puts it this way. Agents can open repos, send patches, review code, run workflows, edit canvases, orchestrate other agents, drop into voice huddles, create channels and pull in whoever needs to see it. The same affordances as a human teammate. The same audit trail. A different keypair.
The foundation matters here. Nostr, the protocol Block chose, solves what the company sees as the core problem in multi-agent systems. Identity. Every participant owns their cryptographic keypair. It belongs to them. Not to whatever platform they happen to use at the moment. That identity is portable. Verifiable. Independent of any vendor. Agents built in Buzz can move across any Nostr-compatible system. Their history and reputation travel with them.
Buzz stays model-agnostic and agent-agnostic too. Teams can bring agents powered by any large language model or framework. Options include tools like Claude Code, Codex or the open-source goose project. They can deploy their own or create new ones. The workspace simply provides the shared environment. It doesn’t dictate what runs behind each agent.
Integration with existing systems forms another key piece. Connect a database. A customer relationship management platform. A codebase. A file system. Agents gain access according to the permissions set by the team. For organizations that self-host Buzz, this delivers complete control over data, relays and agents. Nothing passes through a third party unless deliberately chosen. Block also offers a hosted version at buzz.xyz for teams that prefer managed infrastructure while retaining the protocol’s guarantees.
Open source forms the philosophical core. Block argues that companies currently build agent infrastructure inside closed platforms run by a few large providers. Those platforms set their own rules on access, communication and model usage. The result is fragmentation. Vendor lock-in. Difficulty standardizing operations across an organization. Buzz takes the opposite approach under an Apache 2.0 license. Teams select the models and frameworks that suit them. They avoid lock-in. They own their workspace the way they own their code.
Early Git integration points toward bigger ambitions. The feature remains in initial stages. Yet Block sees a future where open-source communities host entire projects inside Buzz. They discuss ideas there. Review changes. Work alongside agents that help write, test and maintain code. All within one environment. Conversation and code live together. Communities gain a shared home under infrastructure they control.
Observers have taken notice. Coverage appeared quickly across technology publications. TechCrunch reported that Dorsey announced Buzz as a challenger to Slack and GitHub. The platform puts humans and AI agents in the same conversations. It seems to represent more than a personal project from the Block founder. The company behind Square, Cash App, Afterpay and Tidal built it.
The Next Web highlighted the Nostr connection. Dorsey has championed that protocol for years. Buzz arrives as a free, open-source workspace for humans and agents built on it. Reviewers framed the launch as a direct challenge to Slack in the age of AI agents.
Other outlets picked up similar themes. Yahoo Tech noted the effort to reduce reliance on Slack and GitHub. Dorsey himself described Buzz in a post as a new group chat platform for teams of people and agents of all sizes.
The idea of mixing humans and agents in one workspace isn’t entirely new. Tools already let software agents operate inside collaborative environments. What Block emphasizes is linking those agents to cryptographic identities. Similar concepts have surfaced elsewhere. The Open Worldwide Application Security Project proposed an Agent Name Service. Efforts have explored DNS-style discovery for AI agents. Estonia has moved toward digital IDs for them.
Buzz’s approach could appeal to governance and compliance teams. Cryptographic key pairs for both humans and bots create clear audit trails. Every action ties back to a verifiable identity. That setup ticks boxes for organizations worried about accountability as AI use grows inside enterprises.
Yet questions remain. Some analysts wonder whether every company truly needs humans and agents sharing the exact same digital space. Agents often work at speeds far different from people. Text-based chat might require rate limiting for bots. Or risk humans getting drowned out. Git was designed for distributed code edits, pull requests and merges among humans. Extending that model to include autonomous agents introduces complexities.
Block acknowledges the crowded field. Established collaboration tools exist. Large technology companies have their own plans for agent integration. The company bets that one unified substrate can replace the current mix of chat applications, forges, bots, continuous integration dashboards, release tools, search indexes and assorted glue code. Not through magic. But through a single environment where all these elements connect naturally.
The project’s GitHub page strikes a self-aware tone. Yes, it’s another AI-adjacent developer tool. The team apologizes for that. The difference lies in what agents can actually accomplish once inside. They don’t just observe. They act as teammates.
Whether Buzz gains traction will depend on execution and adoption. The open-source nature lowers barriers. Self-hosting options give control-oriented organizations flexibility. Nostr’s decentralized roots align with growing interest in data sovereignty and reduced vendor dependence. Cryptographic identities offer a technical foundation that proprietary platforms might struggle to match without major changes.
Block’s move also reflects broader industry patterns. After years of rapid hiring during the pandemic, many firms have trimmed staff. AI often gets cited as both the cause and the solution. Block took that narrative further than most by linking its cuts so explicitly to the technology. The launch of Buzz suggests the company isn’t just cutting costs. It’s investing in the tools it believes will define the next phase of work.
Early reactions on X mixed curiosity with skepticism. Some users praised the Nostr integration and potential for truly decentralized team collaboration. Others questioned whether developers really want AI agents posting in their Slack channels. Practical tests will decide more than initial announcements.
For now, Buzz is available immediately. Teams can visit buzz.xyz to create workspaces using Block’s hosting. Or they can download the code from GitHub and run their own instances. The company invites developers to bring existing agents or build new ones directly in the platform.
The bet is straightforward. As AI agents become commonplace in business operations, the infrastructure for human-agent collaboration needs to evolve. Block thinks that infrastructure should be open, portable and built on strong identity foundations. If enough organizations agree, Buzz could carve out a significant role. If not, it will join a growing list of ambitious attempts to redefine workplace software in the AI era.
Either way, the experiment reveals much about where the industry is headed. Companies aren’t waiting for AI to mature further before integrating it deeply into daily workflows. They’re building the environments now that assume agents will be full participants. Block, fresh from its own workforce transformation, has placed an early and visible stake in that future.
WebProNews is an iEntry Publication