Jack Dorsey’s Block is rolling out a new workplace platform called Buzz – a free, open-source hub where human employees and AI agents share the same digital space, collaborate, and even hold their own identities.
In an announcement, Dorsey described Buzz as “a new group chat platform for teams of people and agents of all sizes,” explicitly designed to cut Block’s reliance on tools like Slack and GitHub. He emphasized that Buzz is model‑agnostic, decentralized, self‑sovereign, and open source – signaling a deliberate move away from traditional, closed, and centrally controlled SaaS platforms.
A Shared Workspace for Humans and AI
At its core, Buzz is a collaborative environment that blends familiar workplace features with infrastructure specifically tailored to AI:
– Channels and threads for team coordination
– Direct messages for private conversations
– Support for voice and media sharing
– Integrated code repositories
– Automation capabilities that plug directly into AI agents
Unlike classic chat apps that simply bolt on “AI assistants,” Buzz treats agents as first-class participants. They can join channels, respond to threads, access repositories, and trigger workflows just as a human teammate would – but with one crucial twist: they have their own cryptographic identities.
Agents With Their Own Identities
Buzz runs on Nostr, a decentralized protocol originally built for censorship-resistant social networking. In Nostr, every participant – whether a person or a bot – is defined by a keypair: a public key that acts as an identity, and a private key that proves control.
By building on this model, Buzz lets AI agents:
– Hold independent cryptographic identities, not just “accounts” owned by a vendor
– Sign messages and actions, providing a verifiable audit trail
– Operate across different clients and environments while still being cryptographically the same “agent”
– Participate in multiple workspaces or organizations without being locked into a single platform
For organizations, this means interactions are not just a blob of “system messages” coming from an opaque black box. Every action can be traced to a specific agent identity, making governance, compliance, and debugging far more transparent.
Why Nostr Matters for a Work Platform
Choosing Nostr as the underlying protocol is more than a technical curiosity – it reshapes how a work platform can operate.
On traditional platforms, messages, code, and integrations live on centralized servers, regulated by a single company. Permissions, access, and data portability are entirely at that company’s discretion. With Nostr:
– Data is distributed across relays instead of confined to one corporate server
– Identity is controlled by cryptographic keys instead of corporate login systems
– Clients can be swapped, forked, or customized without losing access to the underlying data
– New tools can plug into the same identity and message layer without asking for permission
Buzz rides on this architecture, essentially acting as a client and orchestration layer on top of Nostr. That gives teams a path toward a more open, resilient workplace stack, where switching tools or providers doesn’t mean abandoning their history or relationships.
Challenging Slack and GitHub From a Different Angle
Buzz clearly positions itself as a challenger to Slack and GitHub, but not by copying their business models. Instead of being a proprietary cloud product, Buzz is:
– Free to use
– Open source, enabling anyone to inspect, modify, or host it
– Designed to break lock‑in by building on a public protocol
Slack excels at synchronous team communication; GitHub dominates code collaboration and version control. Buzz takes pieces of both – channels, threads, messaging, and code repositories – and drops them into a single environment that is inherently agent‑friendly. That means:
– Developers can have AI agents that review pull requests, suggest changes, and comment directly in the same space as humans.
– Ops and support teams can rely on agents monitoring channels, triggering alerts, or auto‑responding to common requests.
– Product teams can have research or analysis agents embedded directly into their workflow, rather than living in separate dashboards or interfaces.
It’s less “another chat app” and more of a protocol‑native coordination layer for people and software agents working together.
Rethinking the Operating Model of a Company
By tightly integrating AI agents into everyday workplace tools, Buzz nudges companies toward a different operating model. Instead of tools being passive utilities, Buzz treats them as autonomous collaborators woven into the fabric of communication.
In practice, this could look like:
– Persistent “team agents” that stay with a project from inception to launch, preserving context and history.
– Security and compliance agents that monitor channels and repos in real time, flagging potential issues.
– Documentation agents that auto‑summarize key decisions made in threads and attach them to relevant repos or project spaces.
– Onboarding agents that guide new hires through past discussions, codebases, and processes using the same Buzz interface.
Because agents have their own cryptographic identities, organizations can manage them like digital employees: granting permissions, rotating keys, decommissioning them safely, or sharing them across teams and tools.
Self‑Sovereign Workflows and Data Control
Dorsey’s emphasis on “self‑sovereign” highlights another shift: Buzz is designed so organizations and individuals retain greater control over identities and data.
In a self‑sovereign model:
– Identities are portable: your public key is yours, not tied to a single company.
– Workflows are not trapped inside a closed API – they’re built on open protocols.
– Teams can decide where to host their data, which relays to trust, and what security posture to adopt.
For teams concerned about vendor lock‑in, regulatory requirements, or long‑term data ownership, this approach offers a way to adopt modern collaboration tools without handing over the keys to a single provider.
Model‑Agnostic by Design
A key trait of Buzz is that it is “model‑agnostic.” That means it does not force organizations to commit to a specific AI model or vendor. Instead, Buzz serves as a coordination and identity layer that can plug into different AI backends.
This has several practical advantages:
– Companies can use different models for different tasks (e.g., one for code, another for support, another for analysis).
– They can swap providers if costs, performance, or policies change, without restructuring their entire workspace.
– Internal, self‑hosted models can coexist with external commercial models, all represented as agents inside Buzz.
In effect, Buzz becomes the neutral ground where multiple AI systems and human teams coexist, collaborate, and interact, without the platform itself dictating which intelligence powers them.
How Buzz Can Fit Into Real-World Teams
For real teams, the value of a tool is in what it changes day to day. Buzz’s design opens up a series of practical use cases:
– Engineering teams can run continuous integration bots, review agents, and security scanners as identifiable participants in code channels, making their output easier to track and act on.
– Customer support teams can build agents that summarize tickets, propose responses, or surface historical context, interacting in channels where managers can oversee and correct them.
– Data and analytics teams can have agents that receive raw logs or metrics in a Buzz channel, run analyses, and post visualizations or recommendations.
– Leadership can use agents that digest cross‑team updates, highlight trends, and prepare executive summaries directly within Buzz.
By placing these capabilities next to ordinary human conversations, Buzz reduces context switching and makes AI participation feel less like “using a tool” and more like working with another teammate.
Benefits and Risks of Agent‑Centric Workspaces
The move toward an agent‑centric workspace also introduces new questions and responsibilities.
Potential benefits include:
– Higher automation of repetitive tasks without leaving your main work hub
– Better institutional memory, as agents can preserve and reuse context from past conversations and code
– More experimentation, since teams can spin up new agents cheaply and tie them into existing channels
But organizations also need to think carefully about:
– Security and access control – agents must not overstep their permissions or leak sensitive data.
– Auditability – every action needs to be traceable back to a specific agent identity and key.
– Governance – teams have to decide who can create, modify, or retire agents, and under what processes.
Buzz’s cryptographic identity system on Nostr makes some of this easier – actions can be signed and verified – but the social and organizational layers still need to be designed thoughtfully.
Extending the Protocol Streak
Buzz continues a broader pattern in Dorsey’s work: pushing core online functions – payments, communication, identity, and now workplace collaboration – onto open protocols rather than closed platforms.
By pairing Nostr’s identity and message relay framework with an AI‑friendly, open‑source interface, Buzz suggests a future where:
– Workspaces become interoperable across apps and vendors
– AI agents are treated as persistent protocol‑level participants, not temporary features inside proprietary dashboards
– Teams can innovate on top of a shared, open substrate instead of reinventing the wheel in isolated silos
If this vision gains traction, Slack‑style chat and GitHub‑style code collaboration might become just one of many clients on top of a wider network of identities and agents – rather than standalone destinations.
What Comes Next for Buzz
As a free, open source tool, Buzz’s trajectory will depend heavily on how fast developers and organizations experiment with it. The combination of:
– Nostr‑based identities
– First‑class AI agents
– Integrated communication and code repositories
– Decentralized, self‑sovereign architecture
positions Buzz as an experimental yet ambitious attempt to redefine how teams organize their work in an era where humans and AI increasingly operate side by side.
For companies exploring how to integrate AI into everyday workflows without surrendering control over data, identity, and tools, Buzz offers an early look at what a protocol‑native, agent‑centric workplace might become.