Watch: Words Become Vectors

You just built embeddings by hand. Here is the idea in motion. An embedding turns every word into a point in space, learned so that meaning becomes distance: words that mean similar things land close together, unrelated words land far apart.

Press play. Watch a handful of words sort themselves into clusters, then watch a brand-new word, puppy, land exactly where its meaning belongs, and see why semantic search is just finding the nearest points.