We want you to build the next Git platform on Cloudflare

Original Article ↗ Hacker News Discussion ↗

Cloudflare’s call for developers to build a “next-generation Git platform” on its Workers and Artifacts stack, with a top prize of $25,000 in Cloudflare credits and a trip to San Francisco, is being met with strong skepticism. Many see it as underpaid spec work that deepens dependence on a centralized US-based provider, especially when Git is already decentralized and self-hosted alternatives or tools like Fossil and GitSocial exist. Others question the premise of designing source control around fleets of AI agents, arguing that current productivity gains come from a few agents paired with humans, not from thousands of bots generating hard‑to‑review code.

Competition structure and perceived value

  • Contest: build a “next Git platform for agents” on Cloudflare Workers + Artifacts.
  • Prize: $25k in Cloudflare credits, trip for up to two team members to SF, VIP dinner; source must be permissively licensed; only US/Canada residents eligible.
  • Many see the scope (reimagining global source control for agents) as huge relative to the limited, non-cash reward and geographic/license constraints.
  • Several describe it as “spec work” or marketing to bootstrap adoption of Artifacts with near-free labor.

Reactions to Cloudflare’s offer

  • Some find the compensation insulting or “tone-deaf,” especially credits instead of cash and no full travel coverage.
  • A few note that giving compute/credits as hackathon prizes is common and not inherently bad.
  • One commenter even offers to fund a similar effort independently, suggesting builders should start their own companies instead.

Agents and the “Git for bots” premise

  • Many question the premise of “hundreds or thousands of agents” working on one codebase.
  • Common experience: productivity gains come from 1–3 agents paired with a human, not large autonomous swarms.
  • Concerns that thousands of agents would generate low-quality “slop” and make review and CI unmanageable.
  • Some argue existing git is already adequate for agents; if it’s bad for humans, it’s bad for agents too.

Decentralization, control, and self‑hosting

  • Strong desire for decentralization and avoiding dependence on “billionaire platforms” and Cloudflare as a single point of technical/political failure.
  • Multiple people advocate self-hosting git (bare repos via SSH, homelabs, on-prem hardware) and using S3-compatible storage or R2 buckets directly.
  • Criticism that Cloudflare’s approach further centralizes the ecosystem and inserts a global MITM.

Cloudflare’s broader reputation

  • Split views:
    • Positive: Cloudflare seen as enabling solo devs to run serious apps cheaply and with minimal ops (Workers, D1, R2, DNS, tunnels, etc.). Loss-leading, good DX, “generational run.”
    • Negative: recent products described as vibe-coded, trend-chasing “slop,” and Cloudflare as cheap or overreaching.
  • Some worry Cloudflare could follow Google’s trajectory: generous free tools that later become leverage as incentives change.

Alternative tools and ideas

  • Suggestions that existing systems like Fossil, beads+git, or GitSocial (git+S3/R2) already better support integrated collaboration and decentralization.
  • Ideas that “agentic SCM” might need something closer to multiplayer editing with CRDTs/OT rather than traditional git semantics.

Global access and politics

  • Frustration that a “global” platform runs a competition limited to US/Canada.
  • Some explicitly avoid US-based infrastructure or Cloudflare for political or privacy reasons.