A model like ChatGPT is built from millions of these.
D1 · The Perceptron
One neuron adds up what it hears, compares the total to a bar, and answers yes or no. That is the whole unit, and every model in this course is built from copies of it. Here you will drag its weights and watch a line turn, apply the 1958 learning rule by hand, and prove for yourself that it always finishes on data a line can split. Then you will meet the four inputs it cannot handle, the book that made that failure famous, and the two extra neurons that dissolve it. You finish by training one in Deepbox and then finding the same object, renamed Linear, inside a modern model.
- The Cell That Fires
- Not Every Input Matters
- A Neuron Is a Line
- The Rule That Learns
- It Always Stops, If It Can
- Two Gates Fall, One Does Not
- The First Winter
- Why Activations Exist
- A Neuron Is a Detector
- Two Neurons Beat XOR
- Build One
- Where They Live Today