Perceptrons: wx+b | Build Your Own LLM Workshop #3 [Refreshed]

Perceptrons: WX+B, Weight, Input, Bias, Analog circuits. Part of a Build your own LLM workshop. =========== LINKS Justin's twitter: https://x.com/JustinAngel Workshop overview: https://go.justinangel.ai/substack Deck: https://go.justinangel.ai/deck Google Drive: https://go.justinangel.ai/drive Code exercise: https://go.justinangel.ai/code-3 Excel exercise: https://go.justinangel.ai/excel-3 =========== CHAPTERS 00:00 Workshop Roadmap 00:35 What Is a Perceptron 01:54 Perceptron Math WX Plus B 02:46 Physical Perceptron Demo 05:19 Interactive Web Demo 06:43 Excel Perceptron Playground 09:06 Visualizing Bias and Weight 12:03 Coding a Perceptron 14:28 Wrap Up and Next Steps =============== ABOUT THIS TALK Justin Angel continues the Build Your Own Large Language Model workshop by introducing the perceptron as the smallest unit of a neural network and the day’s deliverable: a working perceptron. He explains Rosenblatt’s 1958 motivation around sensing, storing information, and influencing behavior, then walks through the core formula f(x)=wx+b, showing how input, weight, and bias affect output (including negative values). Using a physical circuit analogy and an interactive demo at gojustinangel.ai/neuralnet, he demonstrates changing parameters to observe linear relationships and a decision boundary around zero. He then recreates the same ideas in Excel with charts and conditional formatting, and in Python via a Perceptron class exercise. He notes frameworks like PyTorch will come later alongside activation functions.