iNaturalist

Original Article ↗ Hacker News Discussion ↗

API and Developer Ecosystem

  • The iNaturalist API is widely praised: no auth for read-only, open CORS, and good usability for demos, tutorials, and hobby projects.
  • People have built species maps, “what’s near me” tools, games (guessing species names), terminal widgets, camera trap dashboards, and desktop utilities around it.
  • A Python client and a desktop app for organizing photos via iNat data are being developed/maintained.
  • Some ask how the map is so fast; replies explain server-side rendering of points (via Elasticsearch) into PNG tiles, with client-side markers only for small sets.

Apps and User Experience

  • Users like both the main iNaturalist app and its companion Seek app; Seek is praised for casual “point-and-ID” use but criticized for a repeated “don’t disturb nature” modal.
  • Some report declining ID quality in Seek and switch to other AI tools.
  • Power users love iNat as a long-term, searchable catalog of their wildlife photos and experiences.
  • There is frustration that the web UI feels stagnant and clunky, mobile apps differ in behavior, and a new rewrite may ship missing features.

Machine Learning Models and Openness

  • Many consider iNaturalist’s computer-vision models best-in-class for plant/organism ID, especially in giving higher-level taxa instead of wrong species.
  • Some criticize that the models and training pipeline are not open, arguing this conflicts with their scientific posture.
  • Others point to emerging open alternatives for plant/insect recognition and discuss training techniques and compute constraints.

Privacy and Geolocation Risks

  • Several comments highlight doxxing risk: dense clusters of observations can reveal home locations, especially when paired with real names or profile photos.
  • iNat offers options to obscure/hide locations, but they’re seen as non-obvious and easy to forget.
  • Suggestions include default coarse precision near home, random offsets, timed “obscured windows,” and user-defined geofences.
  • There is tension between accurate locations for science/collectors and protecting people and sensitive species.

Scientific and Societal Impact

  • Users describe iNaturalist as life-changing for learning local flora/fauna and engaging with nature (likened to “real-life Pokémon”).
  • It’s credited with aiding invasive species tracking, endangered species monitoring, and data sharing to national portals and GBIF.
  • Agencies and museums reportedly monitor iNat data, contact observers for specimens or precise locations, and use it in management and research.
  • Some note edge cases, including alleged misuse to locate poisonous mushrooms.

Related Apps and Ecosystem

  • Other nature ID tools mentioned: Merlin, BirdNET, birdnet-pi/birdnet-go/birda, Flora Incognita, Pl@ntNet, Observation.org/Waarneming.nl, and FOSS options like WhoBird.
  • Merlin is especially praised for bird audio ID and offline models, though some find contributing observations confusing.
  • Discussion also touches on AI’s dependence on large-scale labeled data and the human labor behind it.

Organizational Direction and Community Culture

  • iNaturalist is widely viewed as a “good for the world” citizen-science project with a notably friendly, beginner-welcoming community.
  • One detailed critique worries that focus on growth, new apps, and “Pokémon-like” experiences may be drifting from supporting power users and core citizen-science goals, though this is presented as subjective concern.
  • Some engineering staff note that improved privacy controls and other concerns are “on the radar,” but timelines are unspecified.

Miscellaneous

  • Users share enthusiasm, anecdotes (e.g., animals reacting to recording attempts), and UI feedback (e.g., location behavior on third-party tools).
  • There are job postings at the organization, and some comment positively on the clean homepage design.