eu

The first of a short series by Welch Labs on building a complete neural network in Python. It introduces the data and the architecture of a simple network before later parts cover forward propagation, gradient descent, and backpropagation.

Martin Gorner's Devoxx talk introduces deep learning with TensorFlow, walking through neural networks for MNIST, how to choose an architecture and how to diagnose networks that fail to train.

Andrej Karpathy builds a bigram character-level language model from scratch, introducing torch.Tensor and the framework of training, sampling and loss evaluation. It is the first step toward a GPT-style Transformer in his makemore series.

British Museum curator Irving Finkel teaches how to write cuneiform, with a colleague following along. It is a short, lively introduction to the script.

A reupload of Joel Spolsky's talk teaching spreadsheet users the Excel features they are probably missing, such as keyboard navigation, named ranges, tables and lookups. The original disappeared and this copy comes from the Wayback Machine.

Big Clive shows how to pick up discarded disposable vape devices from the roadside and extract the rechargeable lithium cells inside. He explains why these devices contain the same circuitry as rechargeable ones and how to reuse the cells.

A roughly 10-minute guided Non-Sleep Deep Rest session voiced by neuroscientist Andrew Huberman. It is meant to reduce daytime fatigue, improve focus and aid sleep.

A guide to charging a non-Tesla EV, here a BMW i4, at a Tesla Supercharger in the UK, with do's and don'ts. It shows the process in practice now that the network is open to all cars.

A 1953 US Navy training film on the basic mechanisms in fire control computers: shafts, gears, cams and differentials. It shows how these mechanical parts compute ship and target firing solutions.

Andrej Karpathy builds a GPT from scratch in code, following Attention Is All You Need and GPT-2/3, in about two hours using PyTorch. He connects it to ChatGPT and assumes familiarity with his earlier makemore videos.

Jeremy Howard's 90-minute practical tour of language models. It covers how they work, GPT-4's capabilities and limits, using the OpenAI API, code writing, data analysis and fine-tuning.

An 11-minute guided Wim Hof breathing session built around an expanding and contracting bubble to pace the breath. Wim breathes along with the viewer over ambient sound.

The first episode of Google's Machine Learning Recipes series. It introduces machine learning and builds a first supervised-learning classifier in about six lines of Python using scikit-learn.

Mathologer shows a simple idea, the commutator, that lets anyone who can solve the first layer design their own full Rubik's cube solution. The same idea carries over to other symmetric twisty puzzles such as the Megaminx.

CuriousMarc inspects a newly arrived Teletype Model 19, an early military unit built around a Model 15 keyboard and printer, after an initial rough cleaning. It is the first part of a restoration series and touches on 5-bit Baudot code.

Dodoid gives an in-depth look at an SGI Onyx RealityEngine2, a $250,000 graphics supercomputer from 1993, covering its history, a physical overview, a teardown and some demos and games.

A roughly 90-minute film of John Dobson building a Dobsonian telescope step by step, including hand-grinding the mirror.

Bisqwit builds a Doom or Duke Nukem 3D style software-rendered 3D engine in C from scratch, supporting non-Euclidean geometry. It covers vector rotation and portal rendering with sped-up, tool-assisted coding.

A This Old House segment where Richard Trethewey installs a heat exchanger that uses waste heat from a central air conditioner to heat a swimming pool.