Interactive vector embeddings guide ships with SemanticGuessr game over 43k words
Neon0asis · reddit · 2026-11-05
Isaacus ML engineer Abdur-Rahman Butler released an interactive vector embeddings tutorial covering vectorization, cosine similarity, semantic search, hierarchical retrieval, clustering, zero-shot classification and deduplication, with Python snippets throughout. Its centerpiece game, SemanticGuessr, asks players to draw a loop around a target word on a 2D map of 43,339 English words via vector equations like london - english + french. He compressed the payload from 311 MB to 12.3 MB by reducing embeddings to 512 dims with 4-bit quantization, losing only 6% precision, so it runs fully static in the browser.
More from Apps
- AI modifies video games while they're being played, new demo sparks singularity debate — Confident-Honeydew66 · 2026-11-05
- Radical Pie 1.17 will convert legacy MathType equations after reverse-engineering — EricLengyel · 2026-11-05
- 6 prompts to make ChatGPT write concisely instead of just saying 'shorter' — HeyAmit_ · 2026-11-05
- Orderain ships an MCP server that lets Claude run your ecommerce store on autopilot — Slight_Guest3459 · 2026-11-05
- Dev builds Chrome extension adding a watch-history tab to any YouTube channel — RichardsonDx · 2026-11-05
- Sense launches an always-sensing agent for everything that happens in person — Scobleizer · 2026-11-05