Lecture Notes for YouTube

Complete notes from any YouTube lecture

Not a summary: the whole lecture, part by part, with definitions, formulas, examples and a mind map, each part linked to the second it starts.

Free plan, no card needed

Real notes of a real lecture: open the cheat sheet, jump to a part

STEM

What a neural network is: neurons, layers, weights

3Blue1Brown · 19 min · 6 sections

The structure of the simplest neural network, built up from scratch to recognize handwritten digits: what a neuron holds, how layers pass activations on through weights, biases and the sigmoid, and how the whole thing is one function written with matrices. Learning comes in the next lecture.

Mind map

Neural network
  • Neurons
    • Hold a number from 0 to 1
    • That number is the activation
  • Layers
    • 784 input pixels
    • 2 hidden layers of 16
    • 10 outputs, one per digit
  • Weights and bias
    • Weighted sum of the previous layer
    • Bias sets the threshold
  • Squishing
    • Sigmoid maps to (0, 1)
    • ReLU in modern networks
  • Matrix form
    • σ(Wa+b)\sigma(W\mathbf{a} + \mathbf{b})
    • Fast in libraries
  • Why layers
    • Edges, then patterns, then digits
    • Same idea in speech
1

Why recognizing a digit is hard

A 28×28 pixel image of a 3 is easy for the brain to read, even though the exact pixel values differ completely from one 3 to the next.

Writing a program that takes the 784 pixels and outputs the digit goes from trivial to dauntingly difficult. The goal: build a network that learns to recognize handwritten digits.

Summary Brains recognize a sloppy 3 effortlessly, but writing a program that does it is dauntingly difficult.

2

Neurons and layers

Neuron
A thing that holds a number between 0 and 1.
Activation
The number inside a neuron. For an input neuron it is the grayscale value of its pixel, 0 for black up to 1 for white.

The network in this lecture:

  • Input layer: 784 neurons, one per pixel of the 28×28 image.
  • Hidden layers: two layers of 16 neurons each, an arbitrary choice.
  • Output layer: 10 neurons, one per digit; the brightest one is the network's answer.

Activations in one layer determine the activations of the next, loosely like groups of biological neurons firing cause others to fire.

Summary Neurons hold activations between 0 and 1, arranged in an input layer of 784, two hidden layers of 16 and an output layer of 10.

3

Why a layered structure could work

Digits break into parts:

  • A 9 is a loop up top and a line on the right.
  • An 8 is a loop up top and another loop down low.
  • A loop itself breaks into little edges.

The hope: second-layer neurons pick up little edges, third-layer neurons pick up loops and lines, and the output combines them into digits.

The same layering fits other tasks: speech goes from raw audio to sounds, syllables, words and phrases.

Summary The hope is that layers build up abstraction: pixels into edges, edges into patterns like loops and lines, patterns into digits.

Key takeaways

  • A neuron is just a number between 0 and 1, its activation.
  • Each layer computes σ(Wa+b)\sigma(W\mathbf{a} + \mathbf{b}) from the layer before.
  • Weights say which pattern a neuron picks up; the bias says how strong it must be.
  • The digit network has about 13,000 parameters, and learning means finding them.
  • Modern networks mostly use ReLU instead of the sigmoid.

Glossary

Activation
The number between 0 and 1 a neuron holds.
Bias
A number added to the weighted sum that shifts when the neuron becomes active.
Hidden layer
A layer between the input and the output layer.
ReLU
Rectified linear unit, max⁡(0,a)\max(0, a); easier to train than the sigmoid.
Sigmoid
The logistic curve that squishes any number into the range 0 to 1.
Weight
The number on a connection that scales the activation it carries.

Mentioned

  • Essence of linear algebra, chapter 3 · For a visual understanding of matrix–vector multiplication
  • Lisha Li · PhD in the theory of deep learning, Amplify Partners

Questions for review

  1. Why does the input layer have exactly 784 neurons?
  2. What would the weights of a neuron that detects a horizontal edge look like?
  3. What does the bias change about when a neuron becomes active?

How it works

  1. 1

    Paste a YouTube link

    A lecture, a class, a sermon or a talk. Or just ask ChatGPT or Claude to take notes on a video.

  2. 2

    Pick a style

    Outline, Cornell, STEM with formulas, sermon notes or talk notes.

  3. 3

    Study, print, export

    Read the notes with their mind map, make an exam cheat sheet, or export to PDF, Word or Markdown.

Notes you would have written, if you had the time

Every part, in order

Long lectures are read part by part, so nothing in the middle is skipped. Each part gets a heading, the key points and a short summary.

A mind map of the lecture

See the whole topic at a glance: the main ideas and what belongs to each.

Real formulas

STEM notes keep every equation and derivation, typeset properly, with worked examples step by step.

Cornell, sermon and talk notes

Cue questions beside the notes, scripture references and discussion questions, or a speaker's claims and sources.

Whole courses from a playlist

A course notebook lists every lecture with its topics, and a glossary grows across the course.

Exam cheat sheet

One printable page of definitions, formulas and what the lecturer stressed, from a lecture or a whole course.

Back to the exact moment

Every part and every source links to the second of the video where it starts.

PDF, Word and Markdown

Print it, edit it in Word or Google Docs, or move it into Notion and Obsidian.

Inside your assistant

Ask for notes in ChatGPT or Claude and read them right in the conversation. Your notes are the same everywhere.

Turn your next lecture into notes

Write your first notes free

Just ask your assistant

Once the connector is added, talk to ChatGPT or Claude the way you normally would.

  • Take Cornell notes on https://www.youtube.com/watch?v=…
  • Make a course notebook from this MIT playlist and write the first lectures
  • Make me an exam cheat sheet from my linear algebra notebook
  • Export my notes on last Sunday's sermon to Word

Use it in ChatGPT and Claude

Add Lecture Notes for YouTube as a connector, then say: take notes on this lecture.

Connector address

https://notes.inyour.chat/api/mcp
How to connect

Frequently asked questions

Which videos work?

Any public YouTube video with captions, including auto-generated ones, in any language. Notes can be written in your language even when the video is in another.

Is it free?

Yes. You can write notes for 3 videos a week, each up to 60 minutes long. Reading, the cheat sheet and export of your notes are never limited.

How is this different from a summary?

A summary tells you what a video is about in a few lines. These notes go through the whole lecture part by part and keep the definitions, formulas, examples and references, the way a good student would write them.

Can I use it in ChatGPT and Claude?

Yes. Add Lecture Notes for YouTube as a connector and sign in. Your notes are the same in ChatGPT, in Claude and on this website.

Can I make notes for a whole course?

Yes, with Pro: paste a playlist link and you get a course notebook. The first lectures are written at once and the rest with one click each.

Watch the lecture once, keep the notes forever

Start free