Do we need an AI-native slack?

I think the interesting question isn’t whether we need another Slack, but whether existing collaboration tools were designed for an AI-first workflow. As AI becomes part of everyday work, the way we search, share context, and make decisions probably needs to evolve.

the hardest context to capture isn't explicit decisions but tacit conventions (like "we always return 422 not 400 for validation errors") that live in someone's head and never get written down. does PromptQL surface that kind of ambient knowledge passively as teams work, or does it mostly rely on explicit tagging and thread mentions?

I do not think Slack's problem was ever a missing AI layer. It got bloated because every workflow ended up jammed into a channel and search never caught up, and dropping a pile of agents into that same structure mostly just adds noise that now happens to carry a cryptographic identity.

The more interesting part of Buzz is happening under the chat, where every human and agent gets its own identity and scoped permissions on Nostr. That is the bit I would watch, because the moment an agent posts something it should not have or wanders into the wrong channel, people are going to care about permissions and audit trails way more than whether the threads look nice.

I spend my days on auth infrastructure so I am biased, but agent identity feels like the real battleground here, not the UI.