About MEMEX

An archive of one-source video commentaries on overseas AI/tech talks, written in Japanese.

Etymology

MEMEX is named after the personal knowledge-augmenting device proposed by Vannevar Bush in his 1945 essay "As We May Think." Bush envisioned a microfilm-and-link machine that would let an individual associate and traverse books, notes, and others' documents — a conceptual ancestor of hypertext and the World Wide Web.

The name also plays on accumulating MEMEs (units of memory, cultural genes) flowing through X (Twitter), where this project's primary publishing happens.

Why this site exists

English-speaking AI/tech conferences, university lectures, and podcasts produce a vast amount of high-quality primary information that rarely reaches Japanese-speaking audiences. The videos sit on YouTube, but the language barrier means most Japanese readers stop at "I heard it's interesting."

MEMEX builds a permanent, searchable layer on top of those videos:

  1. A link to the original YouTube video, with auto-translated Japanese subtitles where available
  2. A Japanese commentary covering the thesis, key insights, and timestamped quotes
  3. A profile of the speaker(s) cross-linked to their other talks featured here

Where X posts flow away in a day, MEMEX preserves the same analysis as a permanent, indexable page.

Editorial principles

Only primary sources (videos, official X accounts, official press releases). No hyperbolic language. All quotes appear with timestamps and exact wording. Two layers protect copyright:

  • Videos are linked to YouTube, never re-uploaded
  • Quotes follow fair-use standards (clear primary/secondary relationship, attribution, no modification)

Operation

Independently run as a media operation under the operator's family company. Commercial-compatibility (Creative Commons license compliance, copyright compliance) is a baseline constraint. Photos used on this site are sourced exclusively from Wikimedia Commons under CC BY or CC BY-SA licenses.

Note for English readers

Article bodies are written in Japanese — that is the core value proposition. However, this English UI lets you:

  • Browse the speaker roster with English names and roles (People)
  • See which talks have been featured (Articles)
  • Watch the embedded YouTube video directly on each article page (no Japanese required)

For full English summaries of the talks themselves, the original YouTube videos with auto-translated English captions remain the best path.

Contact

For comments or corrections, reach out via X.