Show HN: Freenet, a peer-to-peer platform for decentralized apps

Original Article ↗ Hacker News Discussion ↗

Architecture & Consistency Model

  • Freenet is described as a decentralized “shared computer” for apps (chat, social, search), not just a storage layer.
  • State is managed per “contract” (WASM module + state). Each contract defines its own commutative merge function, making it akin to a CRDT/CmRDT/CvRDT.
  • Synchronization uses a summary/delta protocol: one peer sends a summary, the other responds with only missing state; duplicates are implicitly ignored.
  • Contracts can implement app-specific strategies (tombstones, truncated logs, time-ordered event logs) to handle edits and deletes.
  • There is no global consensus or total ordering; good for eventually consistent apps, not for transactional or currency use.

Use Cases & Current Apps

  • Demonstrated apps include a group chat system (River), CMS, social network, and Git hosting.
  • Chat uses approximate timestamps plus content-hash tiebreakers and limits to recent messages.

Anonymity, Censorship-Resistance & Harmful Content

  • Old Freenet prioritized built-in anonymity and censorship-resistance; new Freenet aims for pluggable anonymity layers on top of a general platform.
  • Some argue optional anonymity is better architecturally and allows handling illegal content via reputation and filtering.
  • Others see this as abandoning the original core goal (anonymous, censorship-resistant publishing) and worry about name reuse.

Reputation, Incentives & Ghost Keys

  • Long-term vision includes a decentralized reputation system, possibly with “karma” for contributing resources and gating access to high-value services.
  • “Ghost keys” (pseudonymous identities bootstrapped via donations) are proposed as one reputation bootstrap; critics note centralization and suggest crypto-based burns instead.
  • There is skepticism that any sybil-resistant, decentralized reputation without centralized identity is solved; proponents emphasize “raising the cost” and web-of-trust style models.

Old Freenet vs New Freenet / Project Governance

  • The original network continues as “Hyphanet,” focused on anonymity and censorship-resistance.
  • Strong disagreement over reusing the “Freenet” name and redirecting funding/branding to the rewrite; some see it as hijacking, others as a legitimate successor similar to past rewrites.
  • Archived mailing-list discussions about this split, and about culture-war rhetoric, lead several commenters to question project leadership and tone; others dismiss the drama as localized and overblown.

Technical Limitations & Deployment

  • Implemented in Rust; runs as an encrypted UDP overlay with DHT routing and subscription semantics.
  • Currently desktop-focused; mobile support (especially iOS, due to WASM restrictions and background bandwidth) is a major open issue.
  • Node bootstrapping uses centralized “gateway” peers today, with plans to decentralize. Peers track cost/benefit to mitigate flooding and sybil-style abuse, but this is acknowledged as a future hard problem.

Comparisons to Other Systems

  • Compared to: old Freenet/Hyphanet, Tor/I2P, Gnutella/Napster, Braid, blockchains, and Filecoin-like incentive systems.
  • Key distinctions: no global ledger, app-defined merge logic, DHT-based routing, and focus on general decentralized apps rather than just anonymous file sharing or cryptocurrency.