Augmenting Human Intellect is fundamentally about referencing ideas. Often, I’m watching a video and want to share a very specific moment—a concrete clip—inside a Seed document or even a comment. Today, that’s surprisingly hard.

Our concept of Block Types is to define instances of media type representations that may, over time, have different implementations or even come from third-party systems. We already do this with paragraphs, images, code blocks, and videos. But not all video is the same. A Video block is fundamentally different from a Clip block.

Most digital media platforms implicitly recognize this distinction. YouTube separates long-form videos from Shorts. Instagram has Reels. Twitter once had Vine alongside longer videos. These aren’t just differences in length—they are different media primitives with distinct use cases and user expectations.

For Seed, this distinction likely matters at a systems level as well. Long-form video probably requires a dedicated streaming or hosting solution. It doesn’t make sense for a peer-to-peer network to store or replicate gigabytes of video data across users’ machines.

Clips, however, are different. Short, referential video fragments could follow a strategy closer to images within our P2P storage policies: lightweight, addressable, and easy to embed in context.

The UX of clipping a video is also fundamentally different from uploading one long video. Clipping is an act of reference, not publication. It’s something I constantly want to do: isolate a moment, point to it precisely, and weave it into a larger idea.

Conclusion: we should introduce a distinct Block Type for Clip (or Reel / short video), clearly separated from regular Video blocks as a core primitive for reference, thought, and shared understanding.