SuperAPI AI Directory

duo-product

General & other

View repository

Updated 10-15 22:13

About this skill

Duolingo Product — Map of Content

Three of the five handbook principles describe how Duolingo decides what to build and what good looks like. This skill collects the decision rules so a small team can borrow the operating system without copying the product.

This skill is structured as a graph: scan the descriptions below, follow only the [[wikilinks]] you need.

The three decision lenses

  • [[references/take-the-long-view]] — Principle #1: if it helps short-term but hurts long-term, it's not right.
  • [[references/raise-the-bar]] — Principle #2: every feature must be intuitive, delightful, useful, and polished.
  • [[references/ship-it]] — Principle #3: clock-speed, ruthless prioritization, kill what isn't working.

Long-view applied

  • [[references/hundred-year-brand]] — the asset class the long view is investing in.
  • [[references/intuitive-by-default]] — products shouldn't have to explain themselves.
  • [[references/performance-is-access]] — speed on the worst hardware in your user base is an access problem, not a polish item.

Bar-raising applied

  • [[references/dogfooding]] — using the product daily as a quality forcing function.
  • [[references/polish]] — what "polished" actually means in review, with concrete tests.
  • [[references/ownership-clarity]] — every critical task has one owner; ambiguous ownership is the bug.
  • [[references/scope-floor-for-v1]] — a numeric floor below which you don't launch, plus the promise never to call it finished.
  • [[references/quality-tiers-by-reversibility]] — tier effort by audience and how hard the work is to take back.

Ship-it applied

  • [[references/ruthless-prioritization]] — how to cut, not how to fit more in.
  • [[references/kill-criteria-product]] — the rules for stopping a project, not just an experiment.
  • [[references/bounded-deliberation]] — give open decisions an option budget and a time box, and tier advice by experience level.

Scoping and framing what you build

  • [[references/difficulty-with-mitigations]] — publish what's hard about your domain, paired line-by-line with the mechanism that answers it.
  • [[references/teach-the-invariant]] — build the shared core, name the axes of variation, hand off the rest.
  • [[references/fit-the-artifact-to-the-moment]] — when users have no time, ship the flat lookup page, not the guided flow.

Sibling skills

  • [[../duo-retention/SKILL]] — the long view is a retention bet; nodes interlock.
  • [[../duo-experimentation/SKILL]] — Ship It runs on the Show Don't Tell protocol.
  • [[../duo-culture/SKILL]] — Raise the Bar is a hiring problem before it's a product problem.

Sources

  • Duolingo Handbook (2025), Principles #1–3