January 10, 2026
From scrub to grid, cue the roast
Show HN: VAM Seek – 2D video navigation grid, 15KB, zero server load
Tiny 15KB video grid wows devs — then gets roasted for “AI slop” and bad mobile UX
TLDR: A new 15KB tool turns video seeking into a 2D grid with client-side thumbnails and no server load. The top comment cheers the concept but blasts mobile UX, bloated docs, and “AI slop,” sparking a classic debate: bold idea versus launch polish — important for anyone building video experiences.
On Hacker News, a dev dropped VAM Seek, a tiny ~15KB tool that turns a video’s timeline into a clickable 2D thumbnail grid. Instead of guessing with a skinny progress bar, you tap tiles to jump scenes. The slick pitch: client-side thumbnails (your browser does the work), zero server load, and one-line setup. Sounds like a dream for streaming sites and impatient binge-watchers, right?
Then the comments happened. Top voice littlestymaar delivered the mood in one breath: love the idea, hate the execution. They praised the concept — “compelling,” “solves a real use case” — but slammed the mobile experience as “terrible,” called the code and docs “an overly verbose mess,” and dropped the zinger: “you can’t ship that slop mate.” Cue the collective wince. The jab at an “AI generated readme” lit up the age-old fight: prototype with AI? Sure. Ship AI-flavored mess? Hard no.
The drama crystallized into a familiar internet showdown: vision vs polish. Some see a fresh take on video controls that could cut servers a break and help users stop blind scrubbing. Others only see a cool demo begging for a ruthless edit — fewer words, faster thumbs, and a grid that doesn’t crumble on phones. Whether you love it or loathe it, everyone’s watching to see if this flashy grid grows up into a real player.
Key Points
- •VAM Seek provides a 2D thumbnail grid for video navigation as an alternative to the 1D seek bar.
- •The library performs client-side frame extraction, aiming for zero server CPU usage and ~15KB JS size.
- •Features include a 200-frame LRU cache, 60fps marker animation via requestAnimationFrame, and a precise VAM algorithm.
- •Integration is available via CDN or npm, with examples for React, Vue, and vanilla JS and an API for seeking and configuration.
- •A demo backend built with FastAPI requires Python 3.9+ and FFmpeg, offering endpoints for grid config, position-to-timestamp, upload, and health checks.