Jack Dorsey Buzz Hits Slack With Open AI Agent Chat

Jack Dorsey's Buzz is trying to turn workplace chat into the control room for human employees and AI agents, not just another place where people complain about notifications.

Jack Dorsey's Buzz Bets That Slack Missed the AI Agent Moment

Buzz launched Tuesday as a Block-built, open source group chat platform for teams of people and agents, according to TechCrunch. The target is explicit: reduce dependence on Slack and GitHub by putting conversations, agent work, and project coordination into one shared workspace.

Dorsey’s own framing was unusually direct.

“we're launching BUZZ! a new groupchat platform for teams of people and agents of all sizes, built to reduce our dependency on slack and github. model-agnostic, decentralized, self-sovereign, and open source.”

That quote matters because Jack Dorsey isn’t pitching a generic productivity app. He co-founded Twitter and Block, and Block’s website says Buzz was built by the same company behind Square, Cash App, Afterpay, and Tidal. A workplace communication product from that shop deserves more scrutiny than a weekend prototype.

The thesis is clear: Slack assumes chat is a human network with software integrations attached. Jack Dorsey Buzz assumes the next workplace interface will include software agents as visible participants. That’s the real fight. Not chat bubbles. Control.

Inside Buzz's Human and AI Agent Group Chat Model

Buzz’s core move is simple: humans and AI agents share the same workplace conversations. Agents aren’t hidden in side panels or treated only as command-response bots. They sit closer to the center of the workflow.

Based on the supplied product descriptions, Buzz combines channels, threads, direct messages, voice, media sharing, code repositories, and automated workflows. CryptoBriefing reports that each human and agent receives a cryptographic identity with permissions and activity history, while messages, reactions, code changes, workflow actions, and approvals can be stored as signed events on Nostr, a decentralized communication protocol.

That identity model is the sharpest part of the product. If agents are going to review code, run approved workflows, post updates, or contribute to projects, teams need to know which agent acted, under what permissions, and in which context.

A practical Buzz workflow could include agents searching prior discussions, opening repositories, submitting patches, reviewing code, running workflows, editing canvases, and creating channels. Those capabilities come from the additional product details supplied in the source material, not from wishful agent hype.

The risk is just as clear. If agents interrupt too often, misread context, or create new notification streams, Buzz could make work louder. A shared room only helps if it reduces coordination drag. If it adds machine-generated chatter, employees will treat it like another tab to mute.

Buzz's Slack Challenge Runs Through GitHub as Much as Chat

Buzz is positioned against Slack and GitHub, but those are different kinds of systems. Slack owns conversation. GitHub owns code collaboration. Buzz is trying to collapse pieces of both into one workspace.

The Centaur comparison is useful because it exposes the strategic fork. Buzz wants to replace parts of Slack and GitHub. Centaur, as described by Konstantopoulos, wants agents to live inside Slack.

His argument was blunt:

“There’s a lot of room for improvement for agents that live in Slack and can do more work than just coding for teams. In the enterprise setting, this means that you’ll want to self-host for security and control, and you want people to use it in Slack,”

That is the adoption question in one sentence. Do teams want a new agent-native workspace, or do they want better agents inside the tools they already use?

From Twitter Threads to Slack Channels, Dorsey Keeps Returning to Conversation

Dorsey’s career keeps circling back to conversation interfaces. Twitter turned public conversation into a real-time information graph. Buzz aims at private workplace coordination, where decisions, code, and agent actions need a shared record.

The historical pattern is visible even without stretching the source. Workplace tools win when they change who has context. Email copied people into decisions. Chat moved decisions into searchable channels. Git workflows tied code changes to review trails. Buzz is betting that AI agents need to be copied into the work itself, with identity, permissions, and audit history attached.

That’s why the open source angle matters. TechCrunch notes that developers can customize their own Buzz instance because the source code is available. Additional source material says teams can use a hosted version or run Buzz on their own infrastructure, and that the software is available under an Apache 2.0 license.

XOOMAR analysis: this makes Buzz less like a polished SaaS replacement on day one and more like a developer-facing infrastructure bet. That could appeal to technical teams that want control. It could also slow adoption among companies that don’t want to operate another critical collaboration system.

For readers tracking the wider agent-management problem, this connects with our coverage of Meta Warns AI Agents Infrastructure May Crack in 20 Months and the accountability issues raised in 12,000 Agents Haunt Maine ICE Shooting Investigation. Different contexts, same pressure point: once agents act, organizations need traceability.

Employees, CIOs, and Slack Rivals Will Judge Buzz Differently

Employees will judge Buzz by one standard: does it reduce work or create more of it? AI summaries, code review help, workflow execution, and shared search can be useful. But a tool that feels like surveillance, noise, or a management productivity experiment will face resistance fast.

Executives and technical leaders will look elsewhere first:

• Permissions: Which agents can access which channels, repositories, files, and tools?
• Auditability: Can teams reconstruct what an agent did and why?
• Self-hosting: Does running a Buzz instance give enough control to justify operational burden?
• Model choice: Buzz is described as model-agnostic and supports agents powered by different models and frameworks, including Claude Code, Codex, and Block’s Goose, according to the supplied source material.
• Reliability: Can agents perform useful work without creating new review bottlenecks?

The product is early. TechCrunch says Buzz itself admits it is in its “early stages,” and warns it’s probably not a good idea to move an entire team over yet. The free desktop app is available for macOS, Windows, and Linux, and the code has been uploaded to GitHub.

That makes the first audience obvious: AI-native startups, developer teams, and groups already experimenting with autonomous agents. Conservative enterprises will likely wait for proof around security, governance, support, and uptime.

Buzz's Next Test Is Whether AI Agents Become Coworkers or Just Louder Bots

The strongest version of Jack Dorsey Buzz is not “Slack with AI.” It is a workspace where agents can receive tasks, act inside defined permissions, leave an audit trail, and keep project context tied to the original discussion.

The weakest version is easier to imagine: more channels, more notifications, more half-useful bot replies, and more cleanup work for humans.

Buzz’s success won’t depend on beating Slack at chat. It will depend on proving that delegated execution belongs inside the same room where teams make decisions. The evidence to watch is practical: completed integrations, permission controls that teams trust, reliable agent behavior, clear pricing or deployment terms, and signs that early users keep Buzz after the novelty fades.

If Buzz reduces tabs, meetings, and status pings, Dorsey may have found a real wedge. If it merely moves the noise into a new open source shell, Slack and GitHub won’t need to panic.

The Bottom Line

• Buzz signals a shift from workplace chat as messaging software to chat as an operating layer for humans and AI agents.
• Dorsey and Block’s involvement gives the project more weight than a typical productivity startup launch.
• If teams adopt agent-centered collaboration, platforms like Slack and GitHub could face pressure to rethink their core workflows.