References
AI as an Infrastructure
Preface
Part 0: Orientation
1
The Whole Stack in One Pass
2
A Field Map and How to Read This Book
Part I: Foundations and Pretraining
3
Scaling Laws and Compute Allocation
4
Data Curation and Quality
5
Tokenization
6
Transformer Architecture and Its Variants
7
Beyond Dense Transformers: MoE, SSMs, Hybrids
8
Training at Scale: Stability and Distributed Parallelism
Part II: Adaptation and Alignment
9
Supervised Fine-Tuning and PEFT
10
RLHF and Reward Modeling
11
Direct Preference Optimization and the Variant Zoo
12
Synthetic Data and Self-Improvement
Part III: Reasoning and Test-Time Compute
13
Eliciting Reasoning
14
Training Models to Reason
15
The Inference-Time Scaling Paradigm
Part IV: Inference and Serving
16
The Serving Problem
17
Memory and Scheduling
18
Faster Decoding
19
Quantization and Kernels
20
Structured and Long-Context Inference
Part V: Orchestration: Agents, Retrieval, Context
21
Agent Architectures
22
Memory Systems
23
The Harness
24
Multi-Agent Systems
25
RAG and Retrieval
26
Context Engineering
Part VI: Evaluation
27
Benchmarks and Their Discontents
28
Judging and Holistic Evaluation
29
Evaluating Agents and Capabilities
Part VII: Infrastructure and Systems
30
Accelerators and Networking
31
Orchestration and Data Infrastructure
Part VIII: Safety, Interpretability, and Governance
32
Mechanistic Interpretability
33
Scalable Oversight and Control
34
Security and Authorization
Part IX: Ecosystem and Economics
35
The Model Landscape
36
The Tooling Ecosystem
37
The Economics of AI
Summary
References
References
Summary