Reading List
Publications and papers worth reading on large language models and the working practice of building with them. Grouped by category. None of this is paid, none is sponsored.
Foundational papers
-
2017
Attention is All You Need.
Vaswani et al. — the original Transformer paper. Every modern language model traces its architecture to this work.
-
2022
Training language models to follow instructions with human feedback.
Ouyang et al. — the InstructGPT paper, which set the template for RLHF and the modern instruction-following model.
-
2020
Language Models are Few-Shot Learners.
Brown et al. — GPT-3 paper. The first scaling result that hinted at what was coming.
-
2022
Large Language Models are Zero-Shot Reasoners.
Kojima et al. — the "let's think step by step" paper. The origin of chain-of-thought prompting in the form most people know it.
-
2020
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks.
Lewis et al. — the original RAG paper. The architecture under most production LLM systems in 2026.
Industry essays
-
Stratechery
The End of the Beginning.
Ben Thompson's framing of where the AI hardware-software stack ends up. The strategy lens applied to model and infrastructure choices.
-
Latent Space
The Latent Space podcast and newsletter.
Swyx and Alessio's interviews with practitioners building real AI products. The most consistent source of technical signal.
-
Anthropic Research
Anthropic's research blog.
Less marketing, more capability evaluation. Worth reading whenever a new post appears, especially the interpretability work.
-
Simon Willison
Simon Willison's weblog.
The best ongoing daily log of working with LLMs. New tool came out yesterday — Simon already tried it.
-
Andrej Karpathy
Andrej Karpathy's lectures.
The clearest explanations of how these systems actually work. Watch the "Let's build GPT" video if you haven't.
Book-length reads
-
2017
Deep Learning.
Goodfellow, Bengio, and Courville. The standard textbook reference. Older than the transformer revolution but the foundations still hold.
-
2024
Building LLMs for Production.
A practitioner's handbook on shipping LLM features. The operational chapters on monitoring and cost control are the most valuable.
-
2025
DeepLearning.AI short courses.
A library of two-hour focused courses on specific topics. Better than most book chapters and free.