Dat-ecosystem: high level applications built on top of P2P protocols

Commenters debated whether Dat and related P2P stacks have a future beyond hobby projects. Several users (soltanov, janandonly, singularityisne) argued the ecosystem stalls without reliable mobile NAT traversal and incentives for persistent seeding, with janandonly calling Dat mostly dead and singularityisne contrasting Dat’s focus on large payload durability with Nostr’s lightweight relay-and-key model that works only for small messages. Others pointed to new projects and forks - Keet, Quibble, datrs - and saw an undead-but-usable ecosystem: schmoooo insisted Dat is “alive and kicking,” highlighted Rust/C/JS implementations and upcoming datweb to run in browsers, while schmoooo and fsiefken listed apps using iroh-like connectivity and holepunching.

Opinion split sharply over whether recent relay solutions acceptably solve connectivity. fsiefken and proponents of iroh framed NAT traversal as largely solved and pointed to DERP-style fallbacks and specific projects (kith, delta.chat, bones). Critics (soltanov, singularityisne) countered that falling back to DERP or centralized relays concedes true peer-to-peer and reintroduces client-server dependency. Other commenters (prominent660, matheus23) clarified implementation and docs issues - an old, unfinished Iroh blog post and LLM endpoint confusion - and noted Keet is not the same as Cabal, underscoring that debate centers on tradeoffs between decentralization purity and practical connectivity.