PM-specific focus
Every token costs money. Understanding tokenization, context windows, and model tiers lets you set product constraints, write better specs, and own AI feature unit economics.
Deliverable
Model selection memo: use case, quality requirements, cost estimate at scale, chosen model + rationale, risks.
Outcomes
Learning outcomes
Explain context windows to stakeholders
Estimate monthly inference cost
Choose the right model tier
Credible LLM conversations with engineers
Topics
Core topic map
Architecture
Tokens & context
Sampling
Training approaches
Model families
Hallucinations
Multimodal
Benchmarks
Cost & tradeoffs
Tools & frameworks
OpenAI Tokenizer
LMSYS Chatbot Arena
Anthropic pricing calculator
Model cards
Resources
Illustrated Transformer
Karpathy — Intro to LLMs
Anthropic docs
Lilian Weng
State of AI Report
Chapters
Chapter path
Chapter 01 · Active
Understanding Attention Intuition from "The Illustrated Transformer"
Beginner
12–15 min read
Read now →
Chapter 02 · Active
From Transformers to LLMs — The PM Version
Beginner–Intermediate
28–32 min read
Read now →
Chapter 03 · Active
Tokens and Context Windows — The PM Version
Beginner–Intermediate
22–26 min read
Read now →
Chapter 04 · Active
AI Safety, RLHF, and Constitutional AI — The PM Version
Beginner–Intermediate
24–28 min read
Read now →
Chapter 05 · Active
InstructGPT and RLHF — The PM Version
Beginner–Intermediate
22–26 min read
Read now →
Chapter 06 · Active
Fine-Tuning vs Prompting — The PM Version
Beginner–Intermediate
24–28 min read
Read now →
Chapter 07 · Active
Temperature, Top-p, and Sampling — The PM Version
Beginner–Intermediate
22–26 min read
Read now →
Chapter 08 · Active
Why LLMs Hallucinate — The PM Version
Beginner–Intermediate
26–30 min read
Read now →
Chapter 09 · Active
Long Context Window Tradeoffs — The PM Version
Beginner–Intermediate
26–30 min read
Read now →
Chapter 10 · Active
Pre-training vs Fine-tuning vs RLHF — The PM Version
Beginner–Intermediate
22–26 min read
Read now →
Chapter 11 · Active
Hallucinations, Knowledge Cutoffs, and Model Limitations — The PM Version
Beginner–Intermediate
26–30 min read
Read now →