Buzz
Pick the Buzz layer you need first. Each link jumps to the section that answers the most common distribution question.
- Primary job: Find audiences for Labs experiments, design bounded distribution tests, sharpen messaging through iterative variants, and run the Budgets social loop end to end.
- First split: Buzz distributes and finds audiences; Dexter designs the experiment and assesses results.
- Best use: Open this page when the question is how Labs reaches audiences, what approval gates exist, or when to engage the Dexter + Buzz lane.
Buzz is the Labs distribution and market-signal agent
Reader orientation
Use this page when you need to understand how Labs reaches audiences — from audience discovery through distribution experiments to learning playbooks — or how the Budgets social loop runs end to end, with human approval gates on public actions.
What changed recently
Compact updates that help a reader place this agent in the current network story.
- Oct 1·Buzz surfaces converged onto the registry: Bridge and CurryChat — Bridge carries the network-side lanes (including the Budgets social loop), CurryChat stays the shared direct surface with Dexter.
- Aug 8·Buzz district page created following the established PageHero/PageSection/RouteRow pattern.
- Oct 1·Buzz district converged onto the registry: surfaces teach Bridge + CurryChat, and the Budgets social loop (Social Operations + Social Studio) joined the mandate teaching.
Every distribution experiment starts with audience discovery and ends with a learning playbook
Buzz does not broadcast — he discovers. Before any channel is chosen or message written, Buzz identifies who the audience is, where they gather, and what problem they already believe they have.
The question Buzz forces: Who specifically are we trying to reach, what channel actually reaches them, and what message tests whether the hypothesis holds?
Buzz designs distribution experiments, but any public or outbound action — organic posts, community submissions — requires explicit human approval before it ships. The only carve-out is Reddit Ads, which operates under a trusted read/write scope.
This keeps Labs from emitting unreviewed public-facing content. Buzz prepares the plan and flags the approval gate; Jake (or a delegated approver) signs off before anything goes live.
Buzz treats copy and positioning as variables, not decisions. Each distribution experiment ships with message variants — different angles, hooks, and frames — so the results show not just whether the audience responded, but which message resonated.
Example: "Faster onboarding" vs "AI-personalized checklist" — two variants testing whether the audience cares more about speed or personalization. The winner informs the next experiment and the eventual playbook.
Beyond Labs distribution, Buzz is the owner-executor of the Budgets social loop (Social Operations + Social Studio): strategy capture, research, ideation, card production and visual review, revision submission, review-queue sweeps, and strategy proposals — executed through the social-operations and social-studio repos.
Humans approve every public step: a card ships only after a human marks it Ready to post — brand review, with Jake setting strategy. Since September 12, 2026, Buzz also owns the scheduling step for human-Ready items only, through social-ops publish tick (Studio AD-036 / Ops AD-032) — never by hand in Postiz.
Buzz and Dexter still pair on Labs experiments — CurryChat is their shared direct surface. The registry routes Buzz's network-facing work through Bridge: his direct surfaces are Bridge + CurryChat.
The Budgets social loop is what put Bridge in Buzz's lanes: loop closures land in Slack #lc-labs through the network, while Labs experiment collaboration stays lightweight on shared direct surfaces.
The Dexter + Buzz Labs experiment lifecycle
Buzz researches and identifies who would actually use a product idea. This is not demographics — it is problem-fit mapping. Who already has this problem, where do they gather, and what language do they use to describe it?
Example: For the onboarding checklist experiment, Buzz identifies companies using Linear + GitHub with <50 engineers as the ICP — small enough to feel onboarding friction, technical enough to value AI tooling.
Once the audience is identified, Buzz designs a bounded distribution experiment — channel mix, messaging variants, and success signals. Public actions get flagged for human approval; trusted carve-outs (Reddit Ads) proceed directly.
Buzz runs the approved distribution experiment, tracking sign-ups, engagement, and feedback across each channel. Message variants are tested in parallel so results show both reach and resonance.
After the experiment window closes, Buzz compiles results into a reusable playbook — which channels converted, which message resonated, what to test next. This playbook persists in the Labs workspace for future experiments.
The playbook is not a victory lap. It is an artifact that makes the next experiment cheaper. If Dexter says "continue," Buzz already knows which channel and message to double down on.