What's new

Week of November 2, 2026

102 new lessons: Fractions, ratios and rates, Negative numbers and distance from zero, Powers, roots and powers of ten, Formulas: order, substitution and rearranging, Averages and spread, Coordinates, slopes and reading a chart, While loops, Finding bugs when nothing crashes, Functions as values, Sets: unique items and fast lookups, Inheritance and class attributes, Dates, times and time zones, Regular expressions, DataFrames with pandas, Charts with Matplotlib, Project: images are arrays, Python on your own computer, Sending email and text messages, Web pages: HTML and CSS, Scraping web pages, One app, several model providers, Reading math notation as code, Correlation and the shape of data, Comparing groups: t-tests, More tests: groups, categories, relationships, Comparing two models on the same test, Bayesian rates: trusting small numbers, Ranking from head-to-head results, Integrals: totals and areas, Angles and waves, Strange geometry in high dimensions, Curvature and step sizes, Recommender systems, Project: a price model, Support vector machines, Clusters of any shape: DBSCAN and hierarchies, Anomaly detection, Cleaning messy data, Grouping, pivoting and reshaping, Data in pieces: partitions and mergeable summaries, Combining tables with merge, Time series with pandas, Text mining: TF-IDF, search and labels, Location data, Project: an exploratory analysis report, Transfer learning, Autoencoders, Reading C, C++, CUDA and Rust, Binary search trees and traversals, Patterns for coding problems, Project: a spell checker, three ways, Searching for solutions, Project: an unbeatable game opponent, SQL for analysis, Designing a database, Changing a database safely, Config files: YAML and TOML, Pipes and text tools in the shell, JSON on the command line: curl and jq, Shell scripts that fail safely, Project: a word-guessing game, Project: a card game of 21, Project: a drinks kiosk, Project: a flashcard trainer, Project: a price-watch alert bot, Project: Pack Train, an arcade game, Project: a private password-breach check, Packaging your code, Typing in depth, Classes in depth, Make your classes feel built in, Iterators and generators, Context managers, Closures and decorators, Exceptions in depth, Pattern matching, Logging, Command-line programs, Testing with pytest, Code style and tools, Text, bytes and encodings, Binary files, Reading and writing XML, Databases with an ORM, Web pages with Flask, Forms and redirects, Accounts and logins, Project: a trail log web app, APIs with FastAPI, Project: a FastAPI service with accounts, Async HTTP clients, JavaScript for Python programmers, Async JavaScript: promises, await and fetch, A chat front end in the browser, Discovery, capabilities and the two eras, Giving an agent a database, Agents as graphs, Docker Compose, Kubernetes basics, Cloud for AI engineers, Measuring an AI feature from its logs, A/B testing a model change.

314 lessons improved: Your first program, Variables, Numbers and arithmetic, Strings and f-strings, Decisions with if/else, Lists, Loops, Dictionaries, Functions, While loops, Finding bugs when nothing crashes, Functions as values, Comprehensions, Sets: unique items and fast lookups, Errors and exceptions, Modules and imports, Classes and objects, Inheritance and class attributes, Dates, times and time zones, Regular expressions, JSON, Files and CSV data, NumPy arrays, DataFrames with pandas, Charts with Matplotlib, Project: images are arrays, The terminal, Version control with git, Packages and environments, Python on your own computer, How APIs work, Calling APIs from Python, API keys and secrets, Sending email and text messages, Web pages: HTML and CSS, Scraping web pages, Your first LLM call, Capstone: ad-copy reviewer, What an LLM actually does, Tokens and the context window, and 274 more.

Week of October 26, 2026

144 new lessons: Regularization, Bias and variance, Precision, recall and F1, ROC curves and AUC, Cross-validation, Clustering, Dimensionality reduction, The scikit-learn workflow, Capstone: churn prediction, The artificial neuron, Activation functions, Multi-layer networks, The forward pass, Computation graphs, Backpropagation, Build an autograd engine, Weight initialization, The training loop, Mini-batches and SGD, Optimizers, Normalization layers, Dropout, Vanishing and exploding gradients, Debugging training, Residual connections, Convolutional networks, Recurrent networks, Embedding layers, PyTorch tensors, PyTorch autograd and modules, GPUs and compute, Capstone: train a classifier, Language modeling, Tokenization and BPE, A bigram model, Sampling strategies, Neural language models, Attention: the intuition, Queries, keys and values, Scaled dot-product attention, Causal masking, Multi-head attention, Positional encodings, The transformer block, The GPT architecture, Build a tiny GPT, Encoders and embedding models, Pretraining at scale, Scaling laws, Inference and the KV cache, Long context, Mixture of experts, Multimodal transformers, Base models vs assistants, Supervised fine-tuning, Instruction and preference data, RLHF, DPO and direct preference methods, Constitutional AI, Training reasoning models, Prompting vs RAG vs fine-tuning, LoRA, Fine-tune an open model, Distillation, Quantization, Serving models, Speculative decoding, Running open models locally, Benchmarks and their limits, Interpretability, Alignment and safety, Multi-agent systems, Evaluating agents, Reading AI papers, Capstone: a tuned, evaluated assistant, How computers run code, Algorithm complexity, Concurrency and async, Performance and profiling, Core data structures, Recursion, trees and graphs, Searching, sorting and dynamic programming, Testing your code, Debugging systematically, Readable code and refactoring, Types and data modeling, Software design, Databases and SQL, Secure coding, Projects, packaging and CI, Capstone: a command-line app, HTTP from the wire up, API styles, Build an API server from scratch, A web framework, from the inside, Designing a REST API, Validating requests with schemas, A database-backed API, Authentication and authorization, Rate limits, retries and idempotency, Webhooks and events, Documenting APIs with OpenAPI, Running APIs in production, Capstone: a complete API, JSON-RPC 2.0 in depth, Versions and stateless requests, Transports I: stdio, Transports II: Streamable HTTP, Build a server: tools, Resources and prompts, Progress, subscriptions and multi-round-trip requests, Build a client, Hosts: connecting servers to a model, Securing remote servers with OAuth, MCP security, Testing, designing and shipping servers, Capstone: a production MCP server, What an agent harness is, The core loop, built properly, A tool registry for agents, File and search tools, Running code safely, Permissions and approvals, Hooks and skills, Managing the context window, Memory, plans and long tasks, Subagents, Streaming, interrupts and sessions, Evaluation harnesses, Tracing, budgets and guardrails, Capstone: build a coding agent, System design fundamentals, Caches, queues and background jobs, Designing AI systems, Containers and deployment, CI/CD pipelines, Reliability and incidents, Git for teams, Code review, Reading unfamiliar code, Working with AI coding tools, Writing for engineers, Interviews and your portfolio, Capstone: ship a production AI service.

Week of October 19, 2026

99 new lessons: Your first program, Variables, Numbers and arithmetic, Strings and f-strings, Decisions with if/else, Lists, Loops, Dictionaries, Functions, Comprehensions, Errors and exceptions, Modules and imports, Classes and objects, JSON, Files and CSV data, NumPy arrays, The terminal, Version control with git, Packages and environments, How APIs work, Calling APIs from Python, API keys and secrets, Your first LLM call, Capstone: ad-copy reviewer, What an LLM actually does, Tokens and the context window, Prompt fundamentals, System prompts and roles, Examples in prompts, Structured output, Prompt chaining, Choosing a model, Reasoning models and effort, Streaming responses, Conversations and memory, Images and documents, Embeddings, Similarity search, Chunking documents, Vector databases, Retrieval-augmented generation, Making RAG good, Hallucinations and grounding, Tool use, The agent loop, Model Context Protocol, Agent design patterns, LLMs in automations, Prompt injection and security, Evals, LLM-as-judge, Cost and latency, Logging and monitoring, Shipping an LLM app, Capstone: research agent, Functions and graphs, Exponentials and logarithms, Probability basics, Conditional probability and Bayes, Distributions, Mean, variance and spread, The normal distribution, Samples and confidence, A/B tests and significance, Maximum likelihood, Derivatives, Derivative rules, The chain rule, Partial derivatives, Vectors, Vector arithmetic and length, Dot product and cosine similarity, Matrices, Matrix multiplication, Matrices as transformations, Transpose, identity and inverse, Tensors, Eigenvectors (intuition), SVD and PCA, Gradients, Optimization by walking downhill, Softmax, Entropy and surprise, Cross-entropy and KL divergence, What machine learning is, Features and labels, Preparing data, Feature engineering, Train, validation and test sets, Linear regression, Loss functions, Training with gradient descent, Learning rate, Logistic regression, Multiclass classification, Decision trees, Random forests and boosting, Nearest neighbors, Overfitting and underfitting.