I built a faster Notion in Rust

Original Article ↗ Hacker News Discussion ↗

Title and HN Meta

  • Some discussion nitpicks the grammar (“a faster” vs “an faster”) and notes the original title likely had “actually” and was auto-edited by HN, with some pushback on title editing in general.

Product Concept and Performance

  • Many commenters like the focus on speed and the thoughtfulness of the architecture, especially given frustration with sluggish mainstream web apps (Gmail, Notion, Teams, Facebook).
  • Others say Notion already feels fast enough for them and question whether “faster Notion” is a compelling differentiator without new concepts.
  • One person notes it could be a handy out-of-the-box full‑text search system, even beyond its Notion‑like use case.

Authorization, OT, and Scaling Concerns

  • Several technically deep comments question the naivety of assuming Rust and an in-memory auth model will scale cleanly.
  • Critiques focus on:
    • Server workloads often being I/O-bound, making language choice less impactful at scale.
    • The simplicity and possible limitations of the authorization system once real users, updates, sharding, and consistency issues appear.
    • The claim of “Zanzibar-like” behavior: caching permissions in memory doesn’t automatically yield Zanzibar’s consistency guarantees (e.g., New Enemy problem and causal consistency).
  • There are concerns about operational transforms at scale, periodic document rebuilding, and Postgres/TOAST overhead.

Pricing and Business Model

  • Pricing from the blog is cited (~$10/seat; early sponsorship with bonus credits).
  • Some individual users are wary of seat-based SaaS for personal knowledge management, preferring licenses or binaries they can run anywhere.

Open Source, Data Ownership, and Rust

  • Multiple commenters say language choice (Rust) matters less than openness; if it’s closed, they treat it as just another product.
  • Strong themes:
    • Desire for an open-source Notion-like with a robust plugin and schema model, and easy export/sync into a personal knowledge graph.
    • Skepticism of closed tools for long‑term notes, based on painful migrations from proprietary systems.
    • Broader debate over open source economics: some insist on FOSS; others defend closed-source apps that use portable formats and treat “open everything” as structurally favoring hyperscalers.

Alternatives and Ecosystem

  • Many tools are discussed as potential substitutes: AppFlowy, Logseq, Trilium, Thymer, AnyType, Outline, AFFiNE, TiddlyWiki, various personal and experimental projects, and especially Obsidian.
  • Obsidian attracts strong praise (IDE for text, long-form writing, extensibility), but criticism for not being open source and for weaker collaboration out-of-the-box.
  • Several third‑party solutions around Obsidian are mentioned (collaboration plugins, language servers), plus interest in a Rust port of ProseMirror as a reusable library.

Broader Performance & Engineering Culture Rant

  • A long subthread laments modern web bloat: high-end hardware and fast internet still result in slow apps.
  • Explanations raised: incentive misalignment at big tech, feature/promotion culture outranking performance, and lack of testing on lower‑end hardware.

User Feedback and UX Notes

  • Someone hit a JavaScript error on the Outcrop site and comments on suboptimal handling of early-access signups.
  • There’s curiosity about a web/WASM version in addition to the desktop app.