Block’s Bradley Axen launches Buzz, aiming to replace Slack with auditable agent-human workspaces

Block is back in the AI trenches with Buzz, and it has a very specific target: Slack. The pitch, according to the project description and Block’s own framing, is that humans and bots should collaborate in a shared workspace that is “more auditable, sovereign, and secure” than what chat tools can deliver. And this comes on the heels of a brutal reset for the same company, which cut over 4,000 jobs due to AI, per the source. Translation: Block is not treating AI as a side experiment. It is rebuilding its operating model, and it wants the collaboration layer to reflect that.

Buzz is positioned as a free and open source platform based on the decentralized Nostr protocol, so users get cryptographic identity. Block’s developers also describe it as self-hostable, using the word “sovereign,” a term that should ring a bell if you have read Neal Stephenson’s Snow Crash. The stated goal is not just a new chat app. The intent is broader replacement: “Slack, GitHub, and various other communication and collaboration tools,” as the source summarizes. Buzz, in other words, is trying to become the single workspace where both people and agents do the same kind of work humans do, but with verifiable identities attached.

If you are thinking, “Okay, another AI-adjacent developer tool,” the Buzz GitHub repo is ready for you. The repo’s more detailed apology is also more revealing than the headline: it claims that agents inside Buzz can open repositories, 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 comparison that matters is in the repo’s phrasing: “The same affordances as a human teammate, the same audit trail, a different keypair.” That last line is the key. Not just “AI can do things.” The claim is that the actions can be traced to cryptographic identities, instead of anonymous bot behavior buried in logs.

This is not an idea that arrived fully formed on Block’s laptop. The source notes that others have already leaned into identity for agents, including OWASP’s Agent Name Service, “DNS for AI Discovery,” and Estonia’s digital IDs for agents. But Buzz’s approach is framed as governance-friendly in a way that proprietary tools might struggle to match. Block’s head of AI capabilities, Bradley Axen, is quoted saying: “Every company is going to need a place where humans and agents work together,” and adding that the question is whether that place is proprietary or open. He concludes: “We built Buzz because we believe it should be open.” For boards and execs, this matters because it is not only product strategy. It is a platform positioning bet about who controls the “place” where critical work happens.

The article’s commentary also flags a tension executives will recognize immediately. Many incumbents like the concept of “openness,” but when revenue gets involved, barriers tend to follow. The source points to OpenAI selling closed AI, and cites Anthropic’s decision to disallow third-party tools from using Claude subscriptions. In that framing, Buzz is a direct shot at the pattern where early openness later becomes paywalled. Whether Buzz can avoid that fate is a different question, but the reason execs should care is straightforward: agent-human collaboration becomes operational risk the second it touches real systems, credentials, codebases, and approvals.

Technically and organizationally, there is another issue worth chewing on. Keeping humans and bots in the same text-based communication space sounds convenient, but people move at human speed and agents move at machine speed. The source questions the feasibility of shared chat unless bots are rate-limited or mostly talk among themselves. It also points to the precedent that may inform the design: Git exists for many distributed code edits, pull requests, and merges. Those workflows are structured and auditable. Chat is messy and ambiguous. Buzz’s bet, per the source, is that one shared substrate can handle what teams currently “fake with chat, forges, bots, CI dashboards, release tools, search indexes, and a pile of glue code.” The ambition is that collaboration stops being a patchwork and becomes a single system with a consistent audit trail.

Second-order implications are where this gets interesting for executives beyond “cool tool, maybe try it.” If Buzz truly delivers on “cryptographic keypairs” for both humans and agents, it could tighten governance for organizations trying to comply with internal controls and external expectations around traceability. Auditable interactions also make it easier to answer the questions that become unavoidable after an incident: who did what, with what identity, and what changed as a result. That is especially relevant as agent workflows expand from suggestion to action, from drafting to patching, from “chatting” to running workflows. The other implication is competitive: if Buzz’s model catches on, “chat” may be treated as a front-end, while the real collaboration system becomes identity-first and workflow-native. For Block, the calculus looks like a wager that agents need a better home than Slack, and that openness plus cryptographic sovereignty is the differentiator.

Buzz may be “another AI-adjacent developer tool” on paper, but its real stake is the collaboration stack itself. The source ends with a blunt reality check: if large tech companies did not already have designs on this space, and if tools for collaborative agent work did not already exist, Buzz might face fewer hurdles. Still, it is worth a shot, because if the future of work is humans plus agents, the question is not whether an interface changes. It is whether the audit trail, the identity layer, and the governance model change with it.