Week 1 · Further reading

How large language models work, in more depth

The slides told the story from 1948 to today's agents in about 30 minutes. This page has the sources behind each part, with the numbers they support, for when you want to read the original or check a claim. Most links are free to read.

  1. Before the transformer
  2. The transformer
  3. Next-token prediction at scale
  4. Pre-training, then post-training
  5. What next-token prediction explains
  6. From tokens to agents
  7. The falling cost of AI
  8. Benchmarks
  9. Tasks like your homework

1. Before the transformer

2. The transformer

Vaswani et al. (2017), Attention Is All You Need, eight authors at Google Brain and Google Research, dropped recurrence altogether: every word looks directly back at every earlier word, computed as matrix multiplications. That made training parallel on GPUs, let any word reach any earlier word in one step, and works on anything cut into tokens. Each word makes a query, compares it with every earlier word's key, and takes a weighted mix of their values: the three steps on the slide.

Transformers in science

3. Next-token prediction at scale

The model's only output is a probability for every token in its vocabulary. A token is a word or a piece of one; 1,000 tokens is about 750 English words.

SourceThe number
GPT-2 (Radford et al. 2019)a 50,257-token vocabulary
GPT-3 (Brown et al. 2020)175 billion parameters, 96 layers, 96 attention heads, 300 billion training tokens
Scaling laws (Kaplan et al. 2020)loss falls as a smooth power law in model size, data and compute, over more than seven orders of magnitude
Chinchilla (Hoffmann et al. 2022)a 70-billion-parameter model on 1.4 trillion tokens beat much larger models: train on more data, not just more parameters
Llama 3 (Meta, April 2024)over 15 trillion training tokens, a 128,000-token vocabulary; about 85,000 years of reading for one person

4. Pre-training, then post-training

Pre-training teaches what the model knows; post-training, with far less data, teaches how it behaves. Both train the same next-token predictor.

5. What next-token prediction explains

The failure-modes weeks later in the quarter come back to all three.

6. From tokens to agents

An agent is a model in a loop: the model writes tokens; when they are a tool call, the harness runs it and pastes the result back.

WhenWhat
Oct 2022ReAct (Yao et al.): interleave reasoning and actions in one transcript
Jun 13, 2023OpenAI function calling: tool calls as structured output
Nov 25, 2024The Model Context Protocol: one standard way to plug tools into any harness
Feb 24, 2025Claude 3.7 Sonnet and Claude Code: a coding agent in the terminal
Apr 16, 2025OpenAI o3, o4-mini and Codex CLI

Weeks 2 and 3 open this loop up: week 2 reading · week 3 reading.

7. The falling cost of AI

Epoch AI, "The plunging price of thought" (Ament and Burnham, September 22, 2026) is the source of the slide's headline: the cost to reach a fixed score falls about 47% per quarter, roughly 13 times per year, since 2023. Its examples: a GPQA science score that cost about 30 cents a question with o3 in January 2025 cost about 0.04 cents with GPT-5.6 Luna by July 2026, 726 times cheaper in 18 months. For comparison, DNA sequencing got about 1.8 times cheaper per year, computing 1.5, lithium batteries 1.16.

Cautions: the rate differs by task, frontier models stay expensive, and price per question includes hidden reasoning tokens (week 2).

8. Benchmarks

Every "best now" score on the slide came from the Epoch AI Benchmarking Hub on September 29, 2026.

BenchmarkWhat it asksAt launch
GPQA (Nov 2023)448 graduate science questions; experts score 65%, skilled non-experts with the web 34%GPT-4: 39%
SWE-bench Verified (Aug 2024)500 real GitHub issues, screened by 93 developersGPT-4o: 33%
FrontierMath (Nov 2024)unpublished research-level mathunder 2%
Humanity's Last Exam (Jan 2025)2,500 expert questions in over 100 subjectso1: 8%
PostTrainBench (Mar 2026)post-train a base model with one H100 in 10 hours23%, against 51% for the labs' own tuning
OSWorld 2.0 (Jun 2026)108 long computer workflows, about 1.6 hours for a person21%
Terminal-Bench-Science 0.1 (Aug 2026)70 research tasks in the Harbor format, like your homework30%; GPT-6 Astra now 63% (Artificial Analysis)
FrontierMath Erdős (Sep 2026)68 unsolved Erdős problems, stated in LeanGPT-6 Astra solved 2

9. Tasks like your homework

A task is a folder: an instruction, a sealed environment, a reference solution, and tests that write a reward of 1 or 0. The format is Harbor, from Terminal-Bench 2.0. The reference solution must score 1 and an agent that does nothing must score 0. That is how labs measure agents, and how reinforcement learning trains them.