slow networks

Andrej Karpathy's first Neural Networks: Zero to Hero lecture, which builds micrograd, a tiny autograd engine, from scratch in Python. It explains backpropagation and neural network training step by step, assuming only basic Python and some high-school calculus.

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.

The Slow Mo Guys film a TV and screens with a high-speed camera to show how displays build an image. The footage includes CRT-style scanning and how pixels light up over time.

Alex Russell's Chrome Dev Summit 2016 talk on building web experiences that work on slow, flaky networks and low-power devices. It argues developers should focus on the performance of real-world mobile hardware rather than desktop machines.

Daniel Kahneman's Talks at Google lecture on his book Thinking, Fast and Slow, covering System 1 and System 2 thinking and the psychology of judgment and decision making.

Geoffrey Hinton's 2025 Royal Institution Discourse on what AI has learned from biological intelligence and how it might surpass it. He explains neural networks and argues that human language use resembles what backpropagation-trained networks do.

Destin of Smarter Every Day films a cat falling in high-speed slow motion to show how cats right themselves in midair, and introduces the physics behind it.

A Google Tech Talk on Yahoo's 14 rules for faster web pages and the YSlow tool. It covers front-end practices like avoiding @import and fixing ETags, with live analysis of popular sites.

Dolly Parton's song Jolene, slowed down to 33 rpm so that it sounds deeper and more drawn out. The short clip is a well-known example of how pitch-shifting changes a song's mood.

Geoffrey Hinton gives an MIT talk on what is wrong with convolutional neural networks and introduces his capsules project. The description points to the capsule network papers.

The first lecture of Stanford's CS231n course on convolutional neural networks for visual recognition. It surveys the history and key challenges of computer vision, from the Cambrian explosion and Hubel and Wiesel to ImageNet.

Ten minutes of ultra-slow-motion footage of bullet impacts filmed at about one million frames per second by Werner Mehl of Kurzzeit. It includes hollow points in ballistics gel and impacts on glass.

A short National Geographic piece showing a laser system that tracks and kills mosquitoes, filmed in slow motion, as part of anti-malaria efforts. It comes from Intellectual Ventures' 2010 work.

3Blue1Brown's second neural network chapter explains cost functions and gradient descent, showing how a network learns by minimizing error during training.

Fermilab's Don Lincoln explains why light travels slower in water and glass. He knocks down common wrong explanations such as absorption and re-emission, then gives the wave-mechanics reason tied to refractive index.

Smarter Every Day uses high-speed and Schlieren photography to show how a whip crack is a supersonic shockwave formed at the tip. It references academic papers on whip physics.

Steven Johnson's TED talk argues that good ideas come from connected, slow-building hunches and liquid networks, from London coffee houses to Darwin to the modern web, rather than lone Eureka moments.

A short demo video for NVIDIA's StyleGAN paper, which proposes a style-based generator for GANs. It shows face synthesis with style mixing and scale-specific control over attributes like pose and freckles.

The Slow Mo Guys visit a lab with a camera capable of filming at 10 trillion frames per second, fast enough to see light travel. They show the setup and the resulting footage.

CGP Grey compares airplane boarding methods, based on an arXiv paper, and shows why the intuitive back-to-front approach is slow while random or outside-in boarding does better.

3Blue1Brown visualizes Feynman's lecture on the refractive index, explaining why light appears to slow down in glass and why the effect depends on color. It builds the idea from waves produced by oscillating charges in the material.

A university lecture from CS480/680 by Pascal Poupart on attention mechanisms and transformer networks, worked through on a chalkboard.

A Justin Bieber song slowed down 800% with the PaulStretch program, turning a pop track into a long ambient drone.

A guided Wim Hof breathing session of three rounds at a slow pace, building from a 30 second to a 90 second breath hold. An on-screen bubble animation helps you keep the rhythm.

Professor Mike Merrifield of Nottingham explains in a mostly uncut conversation why light travels slower in glass. He treats light as a wave and describes how the material's electrons produce secondary waves that combine to slow it.

OK Go's music video for The One Moment, in collaboration with Morton Salt. It is a choreographed real-time scene that took about 4.2 seconds to film and is shown slowed down to the length of the song.

A Veritasium explainer on how the brain works and how we learn. It is inspired by Kahneman's Thinking, Fast and Slow and covers System 1 and System 2 thinking and why harder learning builds skill.

A short demo showing slow 2D drawing in the Windows 7 GUI compared with XP. It attributes the slowdown to the display driver model lacking hardware accelerated GDI.

Tom Scott visits IEX, a stock exchange that slows down trading with a long coil of fiber. The coil removes the speed advantage of high-frequency traders.

Two Minute Papers covers a paper on speeding up fluid and smoke simulation with convolutional networks. It contrasts the approach with earlier regression forest methods.

Veritasium presents David Sinclair's information theory of aging, which says aging is a loss of epigenetic information, and the evidence that aging can be slowed or reversed. It touches on stressors like fasting, exercise and cold exposure.

Mark Fisher's 2014 talk in Zagreb on the slow cancellation of the future, about cultural stagnation and the sense that popular culture has stopped producing the new.

3Blue1Brown's first chapter on deep learning explains what neurons and layers are and the math underlying a basic neural network, using handwritten digit recognition with visual animations.

SethBling shows MarI/O, a program using neural networks and the NEAT genetic algorithm to learn to play Super Mario World. The description links the source code, the paper and background reading.

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.

Smarter Every Day films Prince Rupert's drops, tempered glass drops, in high-speed slow motion at 130,000 fps. It shows how the drop's heavy internal stress lets the head survive hits while the tail shatters the whole thing.

A two-hour driver's-eye cab ride along the Bernina railway from St. Moritz in Switzerland to Tirano in Italy. The UNESCO-listed line crosses the Alps at high altitude with steep adhesion gradients.

Max Tegmark's MIT CBMM talk on why deep learning works so well. He argues that the laws of physics and the symmetries in physical data make it compressible into small neural networks, tying physics to network structure.

BobbyBroccoli documentary on the rise of Nortel Networks, once the dominant Canadian telecom giant, during the dot-com era. It is part one of a two-part story.

Mark Rober and Jabril build an app that uses machine learning to decode baseball sign sequences. The video also explains neural networks in a simple, accessible way.