My reading list [wip]

I am often asked for recommendations on reading materials for those working with AI/ML and LLMs. Here is a short list of what I read (and re-read!). This is a constantly evolving list. If you have something to recommend, let me know in the comments below!

Blogs / Articles#

A small list of blogs that I follow on LLMs, applied ML, engineering and broader AI trends (in no particular order).

• Eugene Yan — Thoughtful writing on building applied LLM systems by a fellow 🇸🇬. There is even an RSS feed. Also check out applied-llms.org (published mid 2024) which is a great starting point for those newer to working with the brave new world of LLMs.

• Lil’Log — Document my learning notes.

• Simon Willison’s Weblog — Lots of practical advice (+reviews) on using the latest and greatest LLMs, and on engineering in general. A true tinkerer.

• Blog — I work to bring AI into production. I write about AI system design.

• Blog and Notes — This blog focuses more on the research side of things.

• martinfowler.com — A really thoughtful engineering blog, with both short- and long-form posts. List of most recent entries. See the collection of articles on software architecture.

• Latent.Space | Substack — Mix of content on applied AI, ranging from technical to business. Their reading list is a great place to start for those newer to working with LLMs.

• One Useful Thing | Ethan Mollick | Substack — Trying to understand the implications of AI for work, education, and life. By Prof. Ethan Mollick. Click to read One Useful Thing, by Ethan Mollick, a Substack publication with hundreds of thousands of subscribers.

• Interconnects | Nathan Lambert | Substack — A wide range of posts ranging from technical to reflective.

From the AI labs:

• Research — Anthropic is an AI safety and research company that’s working to build reliable, interpretable, and steerable AI systems.

Business:

• Acquired Podcast — This is what I listen to during long commutes, road trips, and when doing chores at home. Note that the episodes are seriously long-form (typically ~3h long!) and require some commitment. These have really opened my eyes to “company-making” and inspired me over the years! Some of my favorite episodes: Epic a.k.a. MyHealthOnline (2025), Ikea (2024), Costco (2023), Amazon and AWS (2022), Nvidia 1 & 2 (2022), TSMC (2021) and a follow-up with Morris Chang (2025), AirBnB (2020), Google Maps (2019), Slack (2019).

Cybersecurity:

• Schneier on Security

Books#

• Artificial Intelligence: A Modern Approach (Stuart Russell and Peter Norvig) – a must-read introduction to AI and agents that is even more pertinent today, even though it was (first) written some 10+ years before LLMs and AI agents became mainstream.

Articles / papers#

This list is not meant to be comprehensive. It is simply my personal bookmark list; articles I frequently share when asked about specific topics. Some of these are from blogs mentioned above. Again, these are not in any particular order.

On working with LLMs#

• Patterns for Building LLM-based Systems and Products (Eugene Yan) [2023] – answers “what should I read to get started building with LLMs?”. This is a couple of years old by now, but the core principles of working with LLMs have not changed (even though models and techniques have evolved tremendously).
• LLM powered autonomous agents (Lilian Weng) [2023] – answers “how should I get started with AI agents?”. Again, this is an evergreen article.
• Defeating Nondeterminism in LLM Inference (Thinking Machines Lab) [Sep 2025] – answers “why are LLMs non-deterministic?”
• Lost in the Middle: How Language Models use Long Contexts (arxiv) [2023] – answers “why can’t you simply stuff everything into a model’s large (e.g. 1M tokens) context window and expect it to work well?”
• Generative Agents: Interactive Simulacra of Human Behavior (arxiv) [2023] – answers “what happens when you try to simulate human society via LLMs?”. This is admittedly not really about working with LLMs, but it does shed some light on what kind of emergent behavior we might expect when we let LLMs run wild.

Design patterns for AI agents#

• OpenAI - A practical guide to building AI agents [2025]

Good prompt design guides and references:

• OpenAI - Guide for function calling
• Anthropic: Writing tools for agents, Building effective agents

Tokenization#

The importance of tokenization is often overlooked by folks working with LLMs. To me, many of the quirks of LLMs come down to tokenization strategies (and trade offs).

• Tokenization in large language models (Sean Trott) [2024] – a gentle introduction to tokenization; interesting discussion on whether morphology really matters with tokenization (i.e. is it important for LLMs to truly understand language rules?).
• Let’s build the GPT tokenizer (Andrej Karpathy) [2024] – a must-watch for understanding tokenization (note: 2h 13min long video). There is also a pretty cool tutorial with code, though it doesn’t capture many of the nuances and insights in Karpathy’s video.

Memory#

• Cognitive Architectures for Language Agents (i.e. CoALA framework) [2024]

LLM architecture#

• SLED - Making LLMs more accurate by using all of their layers (Google) [2025] – could be an interesting approach to reduce hallucinations without any additional external dependencies