Note
What Is a Vector?
A beginner-friendly explanation of vectors as lists of numbers that computers can compare.
Purpose
Explain what a vector is in the context of embeddings and semantic search.
This note keeps the idea simple and avoids going too deep into math.
Key Idea
A vector is a list of numbers.
In AI search, a vector is the number-based version of meaning.
Explanation
In the previous note, I learned that an embedding turns text into numbers.
The vector is the actual list of numbers that comes out of that process.
Simple flow:
Text
↓
Embedding model
↓
Vector
For example, this sentence:
Building a daily learning routine
might become something like:
[0.12, -0.45, 0.88, 0.03, -0.21, ...]
That list of numbers is the vector.
I do not need to understand what every number means yet.
The important idea is:
The vector is the computer-readable version of the meaning.
Simple Analogy
Think of a vector like a profile made of scores.
For example, a movie could be described like this:
Action: 9
Comedy: 2
Drama: 6
Sci-fi: 8
Romance: 1
That can be written as a list of numbers:
[9, 2, 6, 8, 1]
Another movie with a similar pattern is probably similar in style.
A very different pattern probably represents a very different movie.
Embeddings work in a similar way, except the categories are not simple labels like action or comedy.
The embedding model creates many hidden meaning dimensions.
Why Vectors Matter
Computers are good at comparing numbers.
They are not naturally good at comparing human meaning.
So instead of directly comparing two sentences, the system turns both into vectors.
Example:
Question:
How do I stay consistent while learning?
Note:
Building a daily learning routine
The system converts both into vectors:
Question vector → [0.14, -0.42, 0.81, ...]
Note vector → [0.12, -0.45, 0.88, ...]
Then it compares the number patterns.
If the vectors are close, the meanings are probably close.
Connection To Embeddings
Embedding and vector are closely related, but they are not exactly the same idea.
Embedding = the process or representation that turns meaning into numbers
Vector = the actual list of numbers
A simple way to remember it:
The embedding model creates the vector.
The vector stores the meaning pattern.
Semantic search compares vectors.
Connection To Semantic Search
Semantic search works because stored notes and user questions can both be represented as vectors.
Simple workflow:
Stored note
↓
Vector
User question
↓
Vector
Compare vectors
↓
Return closest meaning match
This is how a system can find a useful note even when the user uses different words.
Qurio Example
A Qurio note might be called:
Building a daily learning routine
Later, someone might ask:
How do I stay consistent while learning?
The words are different, but the meaning is close.
If both pieces of text are converted into vectors, Qurio could compare them and return the right note.
Practical Use
The practical takeaway is not that I need to understand advanced vector math right now.
The practical takeaway is that vectors let computers compare meaning using numbers.
This makes vectors one of the basic building blocks behind:
- embeddings
- semantic search
- recommendation systems
- AI memory
- RAG
- document search
Related Content
Takeaways
- A vector is a list of numbers.
- In AI search, a vector can represent meaning.
- Embedding models create vectors from text.
- Similar meanings usually create similar vector patterns.
- Semantic search works by comparing vectors.
Questions To Explore Next
- How does the computer compare two vectors?
- What is similarity search?
- What is cosine similarity?
- What is vector space?
- How are vectors stored?
- How does this connect to RAG?