▶ YouTube Summaries

Theo covers James Garbet's article "The Three Pillars of JavaScript Bloat."

The 3 pillars of JS bloat

1. Older Runtime Support & "Safety"

• Packages like is-string have deep dep trees just to support ES3 (IE6/7) or ancient Node. Almost nobody needs this.
• Hero Devs maintains ancient codebases — once added 60 deps to axe-object-query for Node 0.4 compat, nearly doubling SvelteKit's deps.
• Some packages exist purely to protect against global namespace mutation (e.g., math-intrinsics re-exports Math functions).
• Cross-realm values (iframe vs parent) require workarounds instead of instanceof.

2. Atomic Architecture

• Packages split to absurd granularity: shebang-regex (2 lines) gets 133M downloads/week. slash gets 96M/week. path-key (4KB) gets 158M/week.
• Most are used by exactly ONE other package by the same author — inline code with extra npm overhead.
• Duplicated versions compound the problem.
• Bigger supply chain attack surface.

3. Ponyfills That Overstayed Their Welcome

• Ponyfills stick around forever after native support ships.
• globalThis ponyfill: 49M downloads/week (native since 2019).
• object.entries: 35M/week (native since 2017).
• indexOf: 2.3M/week (native since 2010).

What to do

• E18 Foundation — essential cleanup work, pruning redundant packages.
• knip tool finds unused deps, dead code, unused imports.
• E18 CLI with analyze mode recommends replacements.
• module-replacements project lists modern alternatives.
• npmx.dev for exploring dependency graphs.
• Theo donated $5,000 to E18 (total balance was only $17K).