Ch.2 Resources: Intelligence or Simulation?

Companion resources for Chapter 2. Full text is in the book.

Get the Book on Amazon

Key Papers

Apple: GSM-Symbolic

LLMs solve math by pattern-matching, not reasoning. ICLR 2025.

Read →

DeepSeek R1

RL-trained reasoning without supervised data. Jan 2025.

Read →

ARC-AGI-2

Abstract reasoning benchmark unsolved by any AI as of 2026.

Visit →

Exercise

Ask any LLM a simple logic puzzle. Then change the names and numbers, keeping logic identical. If the answer changes — you found pattern matching. If it stays correct — you found something more interesting.

Discuss

Join #ch02-intelligence on Discord