Venture Board

Six directions B3 could take next.

An exploration board for BitByBit. Each card is a possible venture. Three are researched and scored. The rest are concepts, written down so we can argue about them and kill the weak ones fast.

Researched deep-research done, verdict attached Concept proposal, not yet validated

Comparison · 3 of 6 researched

Only ventures with a verdict are scored. Dots read left to right: favourable · mixed · against.

Axis Estate OpsConditional Agent-Native UI KitNo-go as pitched b3-cf-stack CloudNo-go as pitched
Pain — is it budgeted? High. Labour cost is a live P&L lever. Real, but evidence is vendor-heavy; WTP unproven. Real DX pain — but Cloudflare fixes it for free.
Field occupied? Yes — Agrithmics (local, offline, govt-backed). Yes — Meta, Vercel & shadcn, all at once. Yes — by Cloudflare itself, plus free starters.
Reachable first customer Hard — slow, relationship-driven RPC sales. Unclear — no proven buyer for the guardrail. None — the starter has zero users today.
B3 fit / moat Medium — offline-first + local context. High identity fit, but zero distribution edge. High skill fit; no audience, no moat.
Top risk Incumbent + tight, loss-making budgets. Three giants own distribution; no moat. Cloudflare absorbs this exact layer — 4 buyouts in 19 months.
Where it goes next Competitor DD + one discovery call. Park, or test a narrow shadcn-plugged guardrail. Park, unless B3 commits to the audience grind first.
Vertical SaaSResearched

Estate Ops

Offline-first field-operations software for Sri Lanka's tea, rubber, and coconut estates. Muster to wages, and audit-ready compliance.

Verdict: Conditional. Validate before building. Local incumbent (Agrithmics) already in the field layer.

Open the brief →
Dev ToolsResearched

Agent-Native UI Kit

A component system built so AI-assisted contributors extend it correctly the first time. Slots, not props. Contribution rules as specs.

Verdict: No-go as pitched. Meta, Vercel & shadcn already shipped it. Narrow maybe as a governance layer.

Open the brief →
PlatformResearched

b3-cf-stack Cloud

Managed hosting and paid modules on top of the open-source Cloudflare full-stack starter B3 ships.

Verdict: No-go as pitched. The starter has zero users, and Cloudflare keeps absorbing this exact layer. Audience first, or park it.

Open the brief →
Dev ToolsConcept

Preview & CI Suite

Cloudflare-native pull-request previews and deploy ergonomics, packaged as a product for teams that live on the edge.

Why B3: extends cf-pr-preview-action, a tool B3 already maintains in the open.

Open the concept →
Vertical SaaSConcept

Vertical Ops Platform

Compliance, labour, and records software for underserved Sri Lankan verticals. Estate Ops is the first instance of a repeatable pattern.

Why B3: turns one vertical win into a platform play across paper-heavy SL industries.

Open the concept →
InfrastructureConcept

Edge Agent Infra

Durable, local-first orchestration for AI agents on Cloudflare's edge. State, retries, and scheduling that survive restarts.

Why B3: plays the Workers, Durable Objects, and Agents SDK strength B3 is already deep in.

Open the concept →

How to read this board

Concepts are cheap. Conviction is earned.

The rules

  • Three researched, three proposed. Estate Ops, Agent-Native UI Kit, and b3-cf-stack Cloud have a deep-research pass and a verdict. The rest are arguments, not evidence.
  • Kill fast. A concept earns a full brief only when someone commits to championing it. Most should die here.
  • Promote on demand. Any concept can get the same research treatment as Estate Ops. Say which and it moves up.

What to do next

  • Cull. Cut cards that neither founder wants to defend.
  • Rank. Order the survivors by conviction and reachable first customer.
  • Research the top one. Run the pass, attach a verdict, repeat.