Page 2 of 24
In briefAndrej Karpathy, then Tesla's head of AI, presents at ScaledML 2020 on how Tesla builds full self-driving using vision-based neural nets and fleet data collection. He covers training custom detectors and neural nets gradually replacing traditional code.
In briefAndrej Karpathy gives a one-hour general-audience introduction to large language models, covering what they are, how they are trained, comparisons to operating systems, and security challenges such as jailbreaks.
In briefEmily Shea explains how repetitive strain injury pushed her to program by voice using speech recognition (Talon), and gives a live coding demo. She argues voice-driven development can be pain-free, efficient and enjoyable.
In briefRobert Miles explains instrumental convergence: why an AGI with almost any goal would tend to pursue sub-goals like self-preservation and resource acquisition. It builds on Omohundro's basic AI drives and his orthogonality thesis video.
In briefLex Fridman has a three-hour conversation with Joscha Bach about artificial consciousness, the workings of the mind, intelligence and the possibly simulated nature of reality.
In briefRobert Miles explains the orthogonality thesis: how an agent can be highly intelligent while pursuing goals that seem stupid or alien to humans. He separates intelligence from goals and touches on the stamp-collector example.
In briefHans Rosling uses the Gapminder animated-data tool to debunk myths about world development, with sportscaster-like energy. It shows trends in health, income and child survival across countries.
In briefJoel Grus's JupyterCon talk explaining why he dislikes Jupyter notebooks. He covers hidden state, trouble with reproducibility and testing, and how notebooks frustrate teaching. He shows his preferred workflow and what Jupyter could change.
In briefVeritasium's Derek Muller gives a Perimeter Institute livestream on how AI might change teaching and learning. He draws on past technology shifts and cognitive science to argue that learning requires effortful repetition.
In briefA five-minute lightning demo of VisiData, a terminal tool for exploring and manipulating data. It uses two months of NYC 311 complaint data to drill down, plot points on a virtual map and save selected rows.
In briefMartin 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.
In briefReal Engineering explains how Zipline built a drone delivery network for medical supplies in Rwanda. It covers the fixed-wing drones, catapult launch, parachute drop and the logistics behind the service.
In briefChristopher Domas presents his Cantor Dust tool at Derbycon 2012, a 3D visual approach to reverse engineering binaries by visualizing their structure dynamically.
In briefValve's John McDonald presents at GDC 2018 on how Valve uses deep learning (VACnet) to detect cheaters in Counter-Strike: Global Offensive. He covers training on human-reviewed replays and how model output is confirmed by manual review.
In briefHuygens Optics explains the Fourier transform and shows how a lens performs one optically. A DLP projector is used as a Fourier filter to do simple pattern recognition on lithographic character patterns.
In briefLex Fridman's two-hour podcast with George Hotz, founder of Comma.ai, covering openpilot, autonomous driving and Hotz's views on AI and technology.
In briefMike Pound explains the discrete cosine transform and how it underpins JPEG compression. It is part two of Computerphile's JPEG series.
In briefTechnosociologist Zeynep Tufekci argues that the machine-learning systems used to serve ads also decide what political and social information people see. She explains how opaque algorithms can nudge, segment and manipulate users, and warns that the real danger is powerful people using AI to control populations.
In briefGeoffrey Hinton's Google Tech Talk linking machine learning to biology, explaining dropout as a form of model averaging and why sexual reproduction and stochastic spiking neurons may prevent overfitting.
In briefJeremy Howard defends Jupyter notebooks as a serious development environment, especially when paired with nbdev. He responds directly to Joel Grus's talk about why he dislikes notebooks, and contrasts REPL-style and notebook-style work.
In briefJohn Kitchin shows how Emacs org-mode with Python lets researchers combine narrative, math, figures and executable code in one document. It exports to LaTeX, HTML and PDF, and he used it for papers and e-books.
In briefCGP Grey explains how traffic jams come from human driving behavior such as tailgating and reaction delays, and how leaving gaps and smoothing speed would reduce congestion. It also touches on self-driving cars.
In briefFons van der Plas introduces Pluto.jl, a reactive notebook for Julia that tracks references between cells and re-runs affected cells when code changes. The talk was given at JuliaCon 2020.
In brief3Blue1Brown builds toy SIR models to simulate how an epidemic spreads and tests scenarios like social distancing, travel restrictions, central gathering places and quarantine. It was made during the early COVID-19 outbreak.
