Machine Cognition

Artificial Intelligence

Artificial intelligence refers to computational systems that perform tasks historically requiring human cognition — perception, language, reasoning, planning, and creativity. Modern AI is dominated by deep learning, a family of statistical methods that learn patterns from data.

Key takeaways

Modern AI is overwhelmingly statistical pattern-learning, not symbolic reasoning.
Transformers — introduced in 2017 — power nearly every frontier AI system today.
Scaling laws show predictable capability gains from larger models, more data, and more compute.
AI capabilities now span text, images, audio, video, code, and physical robotics.

What you'll learn

From perceptrons to GPT — how machine learning works, why deep neural networks succeeded, and what large language and multimodal models are actually doing.

Explore the topics

Deep explainers across the field, from foundational concepts to frontier research.

Machine Learning Basics

Supervised, unsupervised, and reinforcement learning explained.

Deep Learning

Why deep neural networks unlocked the modern AI era.

The Transformer Architecture

Attention mechanisms and the architecture behind GPT, Claude, and Gemini.

Large Language Models

How LLMs are trained, what they can do, and where they fail.

Multimodal AI

Systems that fuse vision, language, audio, and action.

Reinforcement Learning

Learning from reward — from AlphaGo to RLHF.

AI Agents

Autonomous systems that plan, use tools, and act.

Scaling Laws & Compute

Why bigger models keep getting better — for now.

Frequently asked questions

What is artificial intelligence?

AI is the field of building computational systems that exhibit behaviors associated with human intelligence — perception, reasoning, language, planning, and learning.

How do large language models work?

LLMs are transformer neural networks trained to predict the next token in vast text corpora. With enough scale, this objective produces remarkably general language and reasoning ability.

What is deep learning?

Deep learning is a subset of machine learning using artificial neural networks with many layers, capable of learning hierarchical representations directly from raw data.

Are AI systems conscious?

There is no scientific evidence that current AI systems are conscious. They are sophisticated statistical models without subjective experience or intrinsic goals.

What is the difference between AI, ML, and deep learning?

AI is the broad field; machine learning is a subfield where systems learn from data; deep learning is a further subfield using deep neural networks.

Glossary

Full glossary

Neural Network: Computational model loosely inspired by biological neurons, learning via weighted connections.
Transformer: Neural architecture using self-attention, foundational to modern LLMs.
Token: Atomic unit of text (word, subword, or character) processed by an LLM.
Parameter: A learned weight in a neural network; frontier models have hundreds of billions to trillions.
Inference: Running a trained model to produce outputs from new inputs.
RLHF: Reinforcement Learning from Human Feedback — used to align model outputs with human preferences.