Embeddings
How AI turns meaning into a map.
An embedding turns a word, sentence or picture into a list of numbers that captures its meaning. Things that mean similar things end up close together, like points on a map. That's how AI knows "nap" and "sleep" are related even though the words look nothing alike.
How it works
- 01Each piece of text is turned into a long list of numbers: its coordinates.
- 02Similar meanings get similar coordinates, so they sit close together.
- 03Software measures the distance between points to find related things.
Sorting the toy box
When you tidy the playroom, you don't sort alphabetically. Cars go with trucks, dolls with the doll stroller, crayons with paper. You group by what things are and how they're used. Embeddings let a computer sort ideas the same way: by meaning, not spelling. "Bedtime", "nap" and "sleep schedule" all land in the same corner.
Where you'll use it
Smarter suggestions
Why your recipe app suggests a soup you'll like after you saved a similar one.
Grouping feedback
Sort 100 customer comments into themes without reading each one first.
Photo search
Type "beach" in your photo app and find beach pictures, even without labels.
Your nap-time experiment
You can't see embeddings directly, but you can see the idea at work. Paste in your messy to-do list:
Here's my messy to-do list: [paste your list, or use this one: buy diapers, email accountant, book pediatrician, birthday gift for Mia, renew car insurance, plan Sunday meals, reply to daycare, order printer ink, call mom, post on Instagram] Group these by meaning into 3–5 categories with clear names. Then explain in one sentence how you decided which items belong together.