Learn AI, deeply.

100 topics. AI that goes deeper on demand.

0

topics

0

interview Q&As

Most viewed

Prompt EngineeringEncoder-Decoder ArchitectureKnowledge DistillationLLM Evaluation

All topics

Personalize your path →
Prompting

Prompt Engineering

“Adding 'take a deep breath' to your prompt measurably improves LLM accuracy on math problems.”

Explore layer paths →

Architectures

Encoder-Decoder Architecture

“The encoder and decoder in transformer models don't share weights, making them two completely separate learned brains.”

Explore layer paths →

Training

Knowledge Distillation

“A tiny student model trained on a giant teacher's guesses often beats one trained on real labels alone.”

Explore layer paths →

Models

LLM Evaluation

“Most LLM benchmarks are now contaminated because the test data was inside the training set.”

Explore layer paths →

Foundations

Activation Functions

“Without activation functions, stacking 100 neural network layers is mathematically identical to having just one.”

Explore layer paths →

Foundations

Large Language Models

“LLMs predict one token at a time, with no memory of their own reasoning between steps.”

Explore layer paths →

Emerging

Multimodal Reasoning

“Models that see images don't merge vision and language — they translate pixels into token-like embeddings first.”

Explore layer paths →

Models

Multimodal Models

“Multimodal models don't translate images into text first — vision and language share the same token space from the start.”

Explore layer paths →

Foundations

What is Artificial Intelligence?

“The term 'Artificial Intelligence' was coined in 1956 before anyone had built a working transistor-based computer.”

Explore layer paths →

Foundations

Context Windows

The maximum amount of text a model can process and "remember" at once

Explore layer paths →

Infrastructure

Working with AI APIs

The practical engineering knowledge needed to build reliable applications on top of LLM APIs

Explore layer paths →

Agents

Retrieval-Augmented Generation

Giving language models access to external knowledge at inference time without retraining

Explore layer paths →

Models

Fine-tuning

“Fine-tuning a model on just 1,000 examples can outperform a model trained on billions of raw tokens.”

Explore layer paths →

Emerging

Computer Use & GUI Agents

AI systems that operate computers by seeing the screen and clicking and typing like a human

Explore layer paths →

Training

Gradient Descent & Optimizers

The algorithms that adjust billions of parameters to minimize loss, from SGD to AdamW

Explore layer paths →

Architectures

Variational Autoencoders

The probabilistic encoder that learns a compressed latent space for generation and interpolation

Explore layer paths →

Agents

AI Agents

“Most AI agents spend more tokens talking to themselves than they do responding to you.”

Explore layer paths →

Deployment

AI Observability & Monitoring

Detecting drift, failure, and degradation in production AI systems before users notice

Explore layer paths →

Society

AI Bias & Fairness

How training data and model design encode societal biases and the technical approaches to measure and reduce them

Explore layer paths →

Training

Regularization Techniques

The tricks that prevent neural networks from memorizing training data instead of learning from it

Explore layer paths →

Training

Scaling Laws

The power-law relationships between model size, data, and compute that predict AI capability

Explore layer paths →

Agents

Prompt Injection & Agent Security

How adversarial inputs hijack agent behavior and the defenses being developed against them

Explore layer paths →

Agents

AI Planning & Task Decomposition

How agents break complex goals into ordered subtasks using techniques like ReAct, MRKL, and Tree of Thoughts

Explore layer paths →

Deployment

Inference Optimization

The engineering techniques that make LLM inference fast, cheap, and scalable in production

Explore layer paths →

Deployment

Edge AI & On-Device Inference

Running AI models locally on phones and laptops without cloud round-trips

Explore layer paths →

Applications

Conversational AI & Chatbots

Building reliable, context-aware chat applications on top of large language models

Explore layer paths →

Emerging

Reasoning Models

Models like o1 and DeepSeek-R1 that think through extended internal chains before answering

Explore layer paths →

Agents

Tool Use & Tool Calling

How agents select and invoke external tools, APIs, and code interpreters to extend their capabilities

Explore layer paths →

Agents

Agent Frameworks

LangChain, LlamaIndex, AutoGen, and CrewAI compared — when to use a framework vs. building your own

Explore layer paths →

Emerging

Long Context & Memory

How modern models handle million-token contexts and why retrieval still beats raw context length

Explore layer paths →

Applications

AI Content Generation

How businesses integrate LLMs into writing, marketing, and creative workflows at scale

Explore layer paths →

Applications

Computer Vision Applications

Real-world uses of vision AI: autonomous vehicles, medical imaging, quality control, and beyond

Explore layer paths →

Applications

AI Coding Assistants

How Copilot, Cursor, and Claude Code work under the hood and how to get the most out of them

Explore layer paths →

Foundations

Diffusion Models

The mathematical framework behind modern AI image and video generation

Explore layer paths →

Emerging

Sparse Autoencoders

The interpretability technique that finds human-readable features hidden inside neural network activations

Explore layer paths →

Applications

Synthetic Data

Using AI to generate training data when real-world data is scarce or privacy-restricted

Explore layer paths →

Infrastructure

AI Infrastructure (GPUs)

The hardware, cloud services, and systems engineering powering AI training and inference at scale

Explore layer paths →

Emerging

Neuromorphic Computing

Brain-inspired hardware architectures designed to run AI far more efficiently than GPUs

Explore layer paths →

Agents

AI Memory Systems

How models retain context across long sessions using short-term and long-term memory architectures

Explore layer paths →

Agents

Multi-Agent Systems

How multiple autonomous AI agents collaborate, debate, and solve complex problems together

Explore layer paths →

Models

Mechanistic Interpretability

Trying to reverse-engineer the black box of neural networks to understand exactly how they think

Explore layer paths →

Ethics

AI Regulation

How governments and institutions are creating rules to govern the development and deployment of AI

Explore layer paths →

Foundations

Neural Networks

The foundational multi-layered architecture inspired by the human brain

Explore layer paths →

Agents

Agent Evaluation & Benchmarking

How to measure whether an autonomous agent actually accomplishes goals reliably and safely

Explore layer paths →

Models

Small Language Models

Highly optimized, specialized models designed to run locally on edge devices

Explore layer paths →

Emerging

Foundation Models for Science

Large pretrained models fine-tuned for genomics, climate, physics, and other scientific domains

Explore layer paths →

Society

Digital Divide & AI Access

Why AI benefits are not evenly distributed and the technical, economic, and policy factors driving inequality

Explore layer paths →

Foundations

Backpropagation

The calculus-based learning engine that powers all modern AI training

Explore layer paths →

Emerging

Model Merging

Combining the weights of independently trained models to create a single stronger model for free

Explore layer paths →

Agents

Agentic Workflows

Orchestrating sequences of LLM calls, tool uses, and decision points into reliable end-to-end pipelines

Explore layer paths →

Foundations

Embeddings

“Two words with opposite meanings can sit closer together in embedding space than two words that mean the same thing.”

Explore layer paths →

Society

AI & the Future of Work

Which jobs automation threatens, which it augments, and how economists and workers are responding

Explore layer paths →

Applications

AI for Data Analysis

Using language and vision models to query, summarize, and reason over structured and unstructured data

Explore layer paths →

Foundations

Machine Learning Fundamentals

How machines learn patterns from data without being explicitly programmed for every case

Explore layer paths →

Prompting

Function Calling

Enabling language models to call external tools and APIs in a structured, type-safe way

Explore layer paths →

Ethics

Hallucinations

When language models generate confident, fluent, but factually incorrect information

Explore layer paths →

Ethics

AI Safety

Ensuring AI systems behave as intended and don't cause unintended harm as capabilities grow

Explore layer paths →

Architectures

Convolutional Neural Networks

The spatial filter-based architecture that gave machines the ability to see

Explore layer paths →

Infrastructure

KV Caching

The critical memory optimization that makes text generation fast and efficient

Explore layer paths →

Infrastructure

PEFT & LoRA

How to fine-tune massive models on consumer GPUs by updating only a tiny fraction of parameters

Explore layer paths →

Infrastructure

AI Data Pipelines

The massive engineering challenge of scraping, cleaning, and preparing internet-scale training data

Explore layer paths →

Society

Model Alignment

The philosophical and technical challenge of ensuring AI systems share human values

Explore layer paths →

Society

Copyright and AI

The legal battles and implications of training generative models on scraped internet data

Explore layer paths →

Infrastructure

Model Quantization

Compressing massive neural networks to run on consumer hardware by reducing precision

Explore layer paths →

Training

Loss Functions

The mathematical objectives that tell a neural network how wrong it is and what to fix

Explore layer paths →

Emerging

World Models

AI systems that build internal simulations of physical reality to plan and predict before acting

Explore layer paths →

Training

Distributed Training

How a single model is trained across thousands of GPUs in parallel using data and tensor parallelism

Explore layer paths →

Deployment

Continuous Batching

The serving technique that processes multiple requests simultaneously to maximize GPU utilization

Explore layer paths →

Agents

Browser Agents & Web Automation

Agents that navigate websites, fill forms, and extract data by seeing and interacting with the browser

Explore layer paths →

Foundations

Chain-of-Thought Reasoning

“Asking a model to 'think step by step' can double its accuracy without changing a single weight.”

Explore layer paths →

Architectures

Graph Neural Networks

Neural networks that operate on graph-structured data like molecules, social networks, and code

Explore layer paths →

Architectures

State Space Models & Mamba

The linear recurrence-based alternative to attention that scales linearly with sequence length

Explore layer paths →

Architectures

Normalizing Flows

Generative models that learn exact likelihood by transforming simple distributions into complex ones

Explore layer paths →

Deployment

MLOps

The practices and tooling for shipping, monitoring, and maintaining ML models reliably at scale

Explore layer paths →

Deployment

Speculative Decoding

Using a small draft model to generate tokens that a large model verifies in one forward pass

Explore layer paths →

Society

AI Ethics

The moral questions raised by increasingly capable AI systems and who is responsible for getting them right

Explore layer paths →

Society

Deepfakes & Synthetic Media

How synthetic media is generated, detected, and regulated in an era of cheap and convincing AI fakes

Explore layer paths →

Society

AI Environmental Impact

The energy, water, and carbon cost of training and running large AI models at scale

Explore layer paths →

Society

AI in Education

Personalized tutoring, academic integrity, and how AI is reshaping teaching and learning

Explore layer paths →

Emerging

AI Video Generation

How models like Sora and Kling generate temporally coherent video from text and images

Explore layer paths →

Emerging

Embodied AI & Robotics

AI systems that perceive and act in the physical world, from warehouse robots to humanoids

Explore layer paths →

Society

AI Consciousness & Sentience

The philosophical and scientific debate over whether AI systems can be conscious, feel, or have moral status

Explore layer paths →

Emerging

AI for Scientific Discovery

How AI is accelerating research in protein folding, drug discovery, materials science, and mathematics

Explore layer paths →

Applications

AI-Powered Search

How retrieval and ranking systems are being rebuilt around semantic understanding and LLM reasoning

Explore layer paths →

Foundations

Attention Mechanism

“Attention doesn't actually read text in order — it processes every word simultaneously against every other word.”

Explore layer paths →

Infrastructure

Vector Databases

Databases optimized for storing and searching high-dimensional embedding vectors at scale

Explore layer paths →

Models

RLHF

“RLHF models learn human preferences from comparisons, never from explicit right-or-wrong labels.”

Explore layer paths →

Models

Open vs Closed Source Models

The trade-offs between publicly available model weights and proprietary API-only models

Explore layer paths →

Foundations

Tokenization

How raw text is chopped into numbers that language models can actually process

Explore layer paths →

Models

Mixture of Experts

Routing inputs to specialized sub-networks to scale parameter count without scaling compute

Explore layer paths →

Models

Vision Transformers

Applying the self-attention mechanism to images instead of text

Explore layer paths →

Models

Generative Adversarial Networks

Two neural networks competing against each other to generate ultra-realistic data

Explore layer paths →

Society

Autonomous Weapons

The ethical dilemma of AI systems making lethal decisions without human intervention

Explore layer paths →

Training

Transfer Learning

How pretrained knowledge is recycled and adapted to new tasks with minimal additional training

Explore layer paths →

Architectures

RNNs & LSTMs

The sequential memory-based networks that dominated NLP before the Transformer era

Explore layer paths →

Emerging

Test-Time Compute

Scaling model intelligence at inference time rather than training time using search and self-verification

Explore layer paths →

Emerging

Constitutional AI & RLAIF

Training AI to critique and revise its own outputs using written principles instead of human labels

Explore layer paths →

Applications

AI in Healthcare

How machine learning is transforming medical imaging, drug discovery, and clinical decision support

Explore layer paths →

Applications

AI in Finance

Fraud detection, algorithmic trading, and risk modeling in the age of machine learning

Explore layer paths →

Foundations

Transformers

“Transformers process every word in a sentence simultaneously, making them fundamentally blind to word order without a clever positional workaround.”

Explore layer paths →

Access restricted

Join the waitlist to unlock ai++

If you already have access as a beta tester, sign in now. Otherwise, request early access with your email.

Request access

Already invited?

Sign in with your authorized account to access the full site.

Sign in

Tip

If you already have an invited email, use it to sign in—this popup will disappear automatically.