deep modules

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.

Andrew Ng presents unsupervised feature learning and deep learning, showing how algorithms can learn representations from unlabeled data instead of hand-designed features. He gives examples in vision and audio, including visualizations of features learned by successive layers.

3Blue1Brown's visual walkthrough of the attention mechanism inside transformers, covering self-attention, multiple heads and cross-attention. It is chapter 6 of his deep learning series.

ArtosisTV's commentary on Game 1 of DeepMind's AlphaStar (Protoss) against top Zerg player Serral at Blizzcon 2019. Artosis analyzes the match with beginner-friendly commentary.

A DeepSec 2010 conference talk by Ralf Philipp Weinmann on attacking the baseband processor of mobile phones rather than the application processor. It covers how open-source GSM base stations make malicious-tower attacks practical against neglected GSM stacks.

3Blue1Brown's visual explanation of how transformers and GPT-style language models work, tracing data through tokens, embeddings and attention.

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

Dave Jones of EEVblog and Richard Stephenson tour the 70m DSS43 dish at the Canberra Deep Space Communication Complex and explain how NASA talks to Voyager 2. It covers the transmitters, receivers and tracking involved.

Rick Houlihan's AWS re:Invent 2018 deep dive on advanced DynamoDB design patterns. It covers index overloading, GSI sharding, composite keys, graph processing and transactions.

Roaring Kitty's July 2020 analysis of GameStop, arguing a bullish deep-value case when the stock was about $4 and short interest exceeded 100%. It covers fundamentals, cash balance and technicals.

Soup Emporium's deep dive argues that Yellowstone is not a ticking timebomb. It goes through uplift and subsidence, earthquakes and what the caldera really is, to rebut supervolcano doom claims.

Sean Carroll's Talks at Google presentation of his book Something Deeply Hidden. He gives an accessible introduction to quantum mechanics and argues for the many-worlds interpretation and the emergence of spacetime.

Quentin Perret of Google gives a technical deep dive into protected KVM (pKVM) on Arm64, a confidential computing approach for CPUs without hardware extensions. It explains how the nVHE hypervisor manages stage-2 MMU access control to protect guest memory.

A 29-minute cinematic timelapse by melodysheep that journeys trillions of years into the future, covering the death of the sun, the end of stars, proton decay, black hole evaporation and the fate of the cosmos.

Rich Hickey's philosophical talk from the first Clojure Conj in 2010. He argues that good design comes from deliberate, away-from-keyboard thinking, and describes how to feed the background mind so it solves problems.

A People Make Games follow-up investigation into Roblox, built on stories from players and creators after the channel's first video. It covers the platform's treatment of young developers and its moderation.

Adam Conover's talk at a millennial marketing conference argues that generations are an unscientific, condescending idea and that millennials do not exist as a real category. The description notes that researcher Jean Twenge later said he misrepresented some of her work.

A long video essay arguing that Koko the gorilla probably could not actually use sign language. It examines the cherry-picked data and the lack of transparency behind the claims, drawing on published critiques.

A conference presentation by Bert Hubert on how companies lose their technological edge and become outsourcing hubs. It argues the same dynamic applies to whole countries and continents, and ends with suggestions for reversing it.

Ali Rahimi's NIPS 2017 Test-of-Time award talk, in which he argues that modern machine learning has become alchemy, working without being rigorously understood. He calls for more rigor and understanding in the field.

A Veritasium video on AlphaFold, DeepMind's protein structure prediction system, and the protein design work it enabled. It features interviews with John Jumper and with David Baker's protein design group and explains why predicting protein folding matters.

Gamers Nexus benchmarks the Ryzen 9 7950X in its BIOS Eco Mode presets (65W, 105W and stock 170W). It measures the performance cost against the power and heat savings.

A long video essay on the claim that chemicals such as atrazine are feminizing frogs. It covers the science, the EPA and the role of chemical companies.

A Microsoft Build 2019 session in which the WSL team explains the new WSL 2 architecture, which runs a real Linux kernel alongside Windows on a hypervisor. It covers faster file performance, Linux file access from Windows and better application support.

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.

Mobileye co-founder and CTO Amnon Shashua explains what goes into sensing for autonomous driving and whether end-to-end deep learning can work for it. He argues for decomposing the problem.

An AWS re:Invent 2019 session for advanced DynamoDB users on design patterns used by large customers, including Amazon retail. It covers GSI sharding, index overloading, graph processing, composite keys and transactional workflows.

A TED2014 interview in which Charlie Rose talks with Google CEO Larry Page about the company's long-term vision, including balloons for internet access and the DeepMind acquisition. Page also discusses data sharing and privacy.

Ali Rahimi accepts the NIPS 2017 Test of Time Award and argues that machine learning has become alchemy, with practice running ahead of understanding.

Leo Dirac explains how transformer models have replaced LSTMs in NLP. He gives background from bag of words through RNNs and LSTMs, then a technical deep dive into multi-headed self-attention and positional encoding, with sample code.

Yannic Kilcher walks through the Attention Is All You Need paper, which introduced the Transformer, a sequence model based solely on attention without recurrence or convolutions. He reviews the architecture and the paper's translation results.

Simon Knowles, co-founder and CTO of Graphcore, talks about "intelligence processors", chips designed specifically for machine learning. The talk was given at the RAAIS 2017 summit and covers Graphcore's architecture.

A Welch Labs explainer on neural scaling laws and the compute-efficient frontier, asking why model loss seems to hit a line it cannot cross. It draws an analogy to an ideal gas law for AI.

A 42-minute CBS interview with Geoffrey Hinton at the Vector Institute in 2023, covering the history of neural networks and the present and future of AI. He discusses AGI timelines and whether language models truly understand.

Pannenkoek2012's commentated tool-assisted Super Mario 64 run collecting the Watch for Rolling Rocks star with half an A press, using parallel universes and QPU movement. The video is marked outdated in favor of a newer version.

Apple security head Ivan Krstic gives a Black Hat 2016 talk on iOS security, covering HomeKit, Auto Unlock and iCloud Keychain. He explains key hierarchies and how data can stay safe even in adversarial clouds.

Tom7 (suckerpinch) builds and evaluates a series of absurd hard drives, including storage made from ping packets across the internet. The video accompanies his paper and mixes deep technical hacking with comedic storytelling.

Jack Diederich's PyCon 2012 talk argues that classes are overused in Python and walks through real code, refactoring away unnecessary classes, exceptions and modules. It popularized the heuristic that a class with two methods, one being __init__, should be a function.

Joe Armstrong, co-inventor of Erlang, gives a funny and opinionated Strange Loop talk on the tangled state of software: build tools, package managers, entropy and the physical limits of computation. He argues for ideas like naming things by content hash and throwing away your first version.

Martin Kleppmann's GOTO Berlin 2016 talk on how collaborative editors, offline apps, and multi-datacenter databases handle concurrent updates. It covers conflict resolution approaches including operational transformation and CRDTs, with a worked look at text CRDTs.