Artificial Intelligence & Machine Learning

Watch

• A.I. ‐ Humanity's Final Invention? ↗Kurzgesagt – In a Nutshell
• A.I. Revolution | Full Documentary | NOVA | PBS ↗NOVA PBS Official
• In the Age of AI (full documentary) | FRONTLINE ↗FRONTLINE PBS | Official
• Large Language Models explained briefly ↗3Blue1Brown
• But what is a neural network? | Deep learning chapter 1 ↗3Blue1Brown
• Will AI outsmart human intelligence? - with 'Godfather of AI' Geoffrey Hinton ↗The Royal Institution

Read

• Life 3.0: Being Human in the Age of Artificial Intelligence ↗A lucid exploration of AI's potential futures, from its immediate societal impact to the profound questions of consciousness and cosmic destiny.Max Tegmark
• AI Superpowers: China, Silicon Valley, and the New World Order ↗An insider's view of the geopolitical race for AI dominance, revealing how machine intelligence is reshaping global economic and technological landscapes.Kai-Fu Lee
• Prediction Machines: The Simple Economics of Artificial Intelligence ↗Deconstructs AI into its economic essence—a prediction technology—offering a clear framework for understanding its business applications and strategic implications.Ajay Agrawal, Joshua Gans, Avi Goldfarb
• Human Compatible: Artificial Intelligence and the Problem of Control ↗A leading AI researcher's urgent call to address the fundamental challenge of aligning advanced AI systems with human values, crucial for safe and beneficial applications.Stuart Russell
• The Unreasonable Effectiveness of DataA seminal argument from Google researchers demonstrating the paramount importance of vast datasets over algorithmic complexity in achieving robust machine learning applications.Alon Halevy, Peter Norvig, Fernando Pereira
• On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? ↗A critical examination of the ethical pitfalls and societal risks inherent in large language models, urging caution in their development and deployment.Emily M. Bender, Timnit Gebru, Angelina McMillan-Major, Margaret Mitchell

Voices to follow

• Andrew Ng ↗A pioneer in deep learning and a leading educator, he has made advanced AI concepts accessible to a global audience.Co-founder of Coursera, former head of Google Brain, Stanford professor
• Fei-Fei Li ↗Instrumental in the development of ImageNet, she champions human-centered AI and its ethical implications, advocating for technology that augments human capabilities.Professor of Computer Science at Stanford University, co-director of Stanford's Institute for Human-Centered AI (HAI)
• Yann LeCun ↗A foundational figure in deep learning, particularly convolutional neural networks, his work has been pivotal in advancing practical AI applications in computer vision.Chief AI Scientist at Meta, Professor at NYU, Turing Award laureate
• Stuart Russell ↗A leading voice on the long-term future of AI, he advocates for the development of beneficial and aligned artificial general intelligence, addressing its profound societal risks.Professor of Computer Science at the University of California, Berkeley, co-author of 'Artificial Intelligence: A Modern Approach'
• Kate Crawford ↗An incisive critic and scholar, she uncovers the societal, political, and environmental costs embedded within AI systems, challenging utopian narratives.Research Professor at USC Annenberg, Senior Principal Researcher at Microsoft Research, author

Glossary

• AlgorithmAn algorithm is a set of step-by-step instructions or rules that a computer follows to solve a problem or complete a task. Think of it like a recipe for a computer. For example, the steps you follow to bake a cake, like "mix flour and sugar, then add eggs," are an algorithm for baking. In AI, an algorithm tells the computer how to learn from data.
• Artificial Intelligence (AI)Artificial Intelligence (AI) is when computers or machines can do tasks that usually require human intelligence, like understanding language, solving problems, or recognizing objects. It's about making machines "smart." For example, when you ask a smart speaker like Alexa or Google Assistant a question, that's AI at work, understanding your voice and finding an answer.
• BiasIn AI, bias happens when a machine learning model makes unfair or incorrect decisions because the data it was trained on was not diverse or representative enough. It reflects prejudices present in the training data. For example, if an AI designed to recommend job candidates was only trained on data from male applicants, it might unfairly favor male candidates, showing bias.
• Computer VisionComputer Vision is a field of AI that enables computers to "see" and understand images and videos, much like humans do. It allows machines to interpret visual information from the world. For example, when your smartphone automatically organizes your photos by recognizing faces or objects like "mountains" or "dogs," that's computer vision at work.
• DataData refers to the raw facts, numbers, images, sounds, or text that computers collect and process. It's the information that machine learning models learn from. For example, all the photos you've ever uploaded to your phone, the songs you listen to, or the words you type are types of data that AI can use.
• Machine Learning (ML)Machine Learning (ML) is a way to teach computers to learn from data without being specifically programmed for every single task. Instead of giving exact instructions, you give the computer lots of examples, and it figures out the rules itself. For example, when a streaming service suggests movies you might like based on what you've watched before, that's machine learning.
• ModelIn machine learning, a model is what the computer creates after it has been trained on data. It's like the learned "brain" or "knowledge" that can then be used to make predictions or decisions. For example, after training an AI with many pictures of different animals, the resulting model is the system that can then identify a new animal in a photo.
• Natural Language Processing (NLP)Natural Language Processing (NLP) is a field of AI that helps computers understand, interpret, and generate human language. It allows machines to communicate with people in a natural way. For example, when you use a translation app to convert text from one language to another, or when your phone corrects your spelling, that's NLP in action.
• Neural NetworkA neural network is a type of machine learning model inspired by the human brain, made of many interconnected "nodes" or "neurons" that process information. It's especially good at finding complex patterns. For example, the technology that allows a self-driving car to recognize traffic signs and pedestrians often uses neural networks.
• PredictionA prediction in AI is when a trained model uses what it has learned from past data to make an educated guess or forecast about new, unseen data. It's not always about the future, but about inferring something. For example, when your email service flags an incoming message as "spam," it's making a prediction based on what it learned from millions of other emails.
• Training (a model)Training a model means feeding a machine learning algorithm a large amount of data so it can learn patterns and relationships within that data. It's like a student studying many examples to understand a subject. For example, to train a model to recognize cats, you would show it thousands of pictures labeled "cat" and "not cat" until it learns what a cat looks like.

Threads 8

Where this connects to other fields, and why it's worth knowing.

• Epistemology Philosophy

  An AI will tell you it's 99% sure of an answer that's completely made up. Engineers now test whether a machine's confidence actually matches how often it's right, and they call it 'calibration.' It's a high-tech version of an ancient puzzle: the difference between truly knowing something and just strongly believing it.

• Social Media & Society Sociology

  Social feeds show you whatever keeps you scrolling, and nothing grips people like outrage. Studies find that every angry, moral word you add to a post makes it spread about a fifth further. So without anyone deciding to, the algorithm keeps breeding a angrier society just by chasing your attention.

• Theories of Justice Law

  When AI predicts things like who might reoffend, people demanded it be 'fair.' But mathematicians proved something wild: you literally cannot satisfy every reasonable definition of fair at the same time. So building 'fair' AI forces a hard choice about which kind of unfairness you're willing to keep.

• How Beliefs About Learning Shape Teaching Education

  Teachers have argued forever: do you drill facts into kids, or let them figure things out themselves? Decades later, people building AI hit the exact same fork, calling it 'supervised' versus 'reinforcement' learning. Two totally separate fields ended up fighting the same fight about how any mind should learn.

• Race & Ethnicity Sociology

  You can tell an AI to ignore someone's race entirely, and it'll still find it anyway, through their zip code, their name, even what they buy. Those clues quietly stand in for race. So a program built to be 'colorblind' can end up rebuilding the same old discrimination, just wearing a data disguise.

• Media Ethics & Representation Media

  Some face-recognition systems badly misread darker-skinned faces, because those faces were barely in the photos the machine learned from. It's the same problem as a TV show where whole groups never appear on screen. Who gets left out of the training data gets left out of how the machine sees the world, hard-coding old blind spots into new technology.

• Endangered and Dying Languages Literature

  People hoped AI translators would rescue dying languages. But an AI only learns a language by reading mountains of text in it, and rare languages barely have any. So the tongues that most need saving are the ones AI quietly ignores, pushing them even closer to vanishing.

• Utilitarianism vs Kantian Ethics Philosophy

  A self-driving car about to crash may have to 'choose' who gets hurt, and a programmer has to write that choice into code before the car ever hits the road. So the old classroom debate about right and wrong, save the most people or never cross certain lines, suddenly becomes an actual engineering spec someone has to type out.