Here's a sentence nobody was expecting to type in 2026: Jack Dorsey is gunning for Slack. Not with a pivot, not with a feature, but with a whole new open-source group chat platform called Buzz—built explicitly for teams where the headcount includes both humans and AI agents.

If that sounds like a niche product chasing a very specific 2026 problem, you're not wrong. But you're also not entirely right.

What Buzz Actually Is (And Isn't)

Buzz, which launched on July 21, 2026, is a desktop chat platform that looks—let's be honest—a lot like Slack. Same channel-based layout, same general vibe. But the key differentiator is that AI agents are first-class citizens here, sitting in channels alongside their human coworkers rather than bolted on as an afterthought via a janky third-party integration.

It also folds in GitHub project management, which means your engineering team could theoretically run code reviews, task coordination, and agent-assisted workflows all from one window. Whether that consolidation is genuinely useful or just feature bloat dressed up in a bee emoji is the real question.

Dorsey described Buzz on X as:

"model-agnostic, decentralized, self-sovereign, and open source."

That's a sentence with four buzzwords (pun intended) that each carry real technical weight. Model-agnostic means you're not locked into OpenAI or Anthropic's APIs—you can plug in whatever agent backend your team prefers. Decentralized and self-sovereign point toward a self-hosted architecture where your data doesn't live on Block's servers by default. Open source means the code is on GitHub right now and anyone can fork it, audit it, or extend it.

Why This Is Actually Interesting (Despite the Hype Reflex)

The enterprise AI agent problem is real and genuinely messy. Most teams running agentic workflows in 2026 are stitching together Slack, GitHub, Linear, Notion, and whatever model API is flavor of the week—with agents living in awkward sidecars, firing off notifications that interrupt humans at inopportune moments, and generally existing in a different conversational layer than the actual work.

Buzz's pitch is essentially: what if the agents just lived in the same room? No separate dashboard, no webhook spaghetti, no "the bot posted in #general again." That's a coherent problem statement, and it's not one Slack has elegantly solved.

Because it's open source, teams with specific workflows can build their own extensions without waiting for a product roadmap that may or may not care about their use case. For security-conscious enterprises, self-hosting means your agent conversations and proprietary data stay on infrastructure you control—a non-trivial advantage over SaaS-hosted alternatives.

The Competitive Landscape Is Already Getting Crowded

Dorsey isn't the only one who spotted this gap. Paradigm's CTO Georgios Konstantopoulos shipped a similar open-source project called Centaur back in May 2026—a self-hosted "virtual employee" that integrates directly into existing Slack workspaces rather than replacing them. His framing was different: keep the tool your team already uses, just add agents that can do more than write code.

These two approaches represent a genuine philosophical fork in the road:

• Buzz: Replace the platform entirely; build something AI-native from the ground up.
• Centaur: Don't fight Slack's network effects; enhance the tool people are already in every day.

Both bets have merit, and both carry real risk. Greenfield adoption is hard. So is convincing enterprises to trust a new self-hosted agent layer with sensitive workflows.

What's Missing From the Press Release

Let's talk about what Dorsey's X post conveniently didn't mention.

• Maturity: Buzz's own website admits it's in "early stages." That's code for "please don't migrate your 500-person engineering team onto this yet."
• Agent orchestration complexity: Putting agents in a chat window is the easy part. Managing context windows, tool-call permissions, agent-to-agent communication, and failure states in a group chat environment is genuinely hard engineering. The docs will tell us how serious this is.
• Slack's moat: Slack has integrations with basically every enterprise tool on Earth, years of workflow inertia, and Salesforce's balance sheet behind it. Switching costs are real and painful.
• Block's track record with consumer products: Square and Cash App are legitimate. Tidal, however, is a cautionary tale about passion projects from founders with divided attention.

Hot Take

The real insight here isn't about Buzz specifically—it's about the category Buzz is trying to define. Every major collaboration platform is going to need a coherent story for how AI agents participate in team communication within the next 18 months. Slack knows this; they've been bolting Agentforce integrations onto their product with varying degrees of elegance. Microsoft Teams is doing the same with Copilot.

Buzz and Centaur are essentially pressure-testing whether a greenfield, agent-first design beats incremental retrofitting of legacy platforms. My prediction: the open-source, self-hosted angle will find a devoted niche—particularly among developer-heavy teams and crypto/web3 shops who already share Dorsey's ideological priors about decentralization. But displacing Slack at the enterprise level? That's a different mountain, and Block's divided attention across Square, Cash App, and now Buzz is a legitimate operational concern.

Worth watching. Not worth migrating your team to. Yet.

What Should You Do Right Now?

If you're building an AI-native product and your team doesn't have years of Slack workflow debt, Buzz is worth a weekend of exploration—the desktop app is free and available for macOS, Windows, and Linux today. If you're in an established enterprise, bookmark it and check back in six months when the rough edges are sanded down.

Either way, the question of where AI agents live in your team's daily workflow is no longer hypothetical. Someone on your team should own an answer to that before it becomes a chaotic mess of competing integrations.

What Do You Think?

Is the future of team collaboration a purpose-built, agent-native platform like Buzz—or will Slack and Teams just absorb this use case with integrations before any challenger can gain traction? Drop your take in the comments.