Collection of some of my favorite things: blogs, books, papers, etc. Unordered. I’ve bolded some of my favorites, and italics means work in progress.
Blogs
I love Ludicity! Nikhil’s prose is fantastic, and he writes with great candor, which is very refreshing. He’s super nice too. Suitable only for audiences with a sense of humour.
Statistical Modeling, Causal Inference, and Social Science
Gelman et al. speak thoughtfully on various stats/computing/science topics. Lots of very interesting posts–most go way over my head, but I always enjoy them.
Interesting articles about computing. Vicki is clearly very knowledgeable.
Jakob often write about Julia, and programming in general. They seem to know a lot about computers and write in-depth content about Julia which I really enjoy.
I like Might’s posts on grad school and functional programming especially–his productivity posts are very interesting too.
Evans often writes about the terminal, and I find it fascinating. Posts are very clear and enjoyable to read.
I’ve only read the debugger series which were very cogent. Certainly something I’ll reference again so I can properly understand debuggers.
The extrovert post and the abroad post were very well written, but I haven’t had the chance to read beyond that.
Papers
Evaluating the Moral Beliefs Encoded in LLMs
I got to see most of Claudia’s talk at JSM 2024, and the paper was just as intriguing. I would be curious to see how modern models perform, especially reasoning models.
I just learnt that I’ve only read the original essay, and that Lockhart further wrote a short book, so I need to read that. The essay itself is very interesting, mirroring a lot of the qualms I’ve maintained about schooling–viz. a focus on answers instead of solutions, the “destination instead of the journey” to mangle a trite phrase. It is a scathing critique of mathematical education, and well worth a read.
Large Language Diffusion Models
These seem super duper cool. I don’t think the transformer (nor LLMs) are the “final form” of “AI”, but I don’t know if this would be the next step. It would be very cool if a bunch of other brilliant folk work on this and we see some rapid progress. I found this through Tim Kellogg’s blog and this article was a splendid introduction to diffusion.
A Tutorial Introduction to the Lambda Calculus
Lambda calculus, and functional programming on a whole, are some of the coolest ideas in CS in my opinion. I really enjoyed reading this paper, seeing Church Numbers, working out a few expressions because I couldn’t believe they actually worked–lambda calculus provided a massive paradigm shift, and this paper was a great way to start. I read an older version but linked to a more recent one. I technically haven’t read that one yet, but I’d hazard that it’s similar.
Work In Progress
Data Science at the Singularity
What are the most important statistical ideas of the past 50 years?
Statistical Modeling: The Two Cultures
Beyond Power Calculations: Assessing Type S (Sign) and Type M (Magnitude) Errors
What Should We Do Differently in STAT 101?
Textbooks
The R book. Hadley writes very clearly about how R works, and I’ve learnt the most about R through this and random experimentation. Not super applied–that would be more R4DS. I have some very rough notes on GitHub when I read most of Adv R.
Category Theory for Programmers (PDF)
Absolutely fantastic. I’m interested in functional programming, and this book explains some FP concepts in a rigorous and cogent manner. I’m working through this slowly and really enjoying it! Notes on GitHub.
I’ve not finished this, in part due to feeling like I’m missing parts of the lessons since I don’t know C++ yet. Zimmerman talks about how to work in an existing codebase/work collaboratively as a computer engineer. I’m finding it very helpful, and all the rules so far seem to make sense in a team context. I should start implementing some of these ideas even in solo work though, like the Rule #4: “Generalization Takes Three Examples”.
Work In Progress
Computer Age Statistical Inference
Monte Carlo theory, methods and examples
Linear Algebra Done Right
Regression and Other Stories
OCaml Programming: Correct + Efficient + Beautiful
Software
I’m not sure how I found Sioyek, but it has quickly become one of my favorite pieces of software. Reading papers/books in a traditional PDF reader (e.g. Okular) works, but Sioyek is clearly built for this content and it shows. There are numerous quality of life features which make interacting with a PDF nicer, such as automatic table of contents, highlighting, bookmarks, “portals”–which are essentially picture-in-picture PDFs. The website has video demonstrations which better explain the features. Highly recommend.
Work In Progress
Positron
Librewolf/Ironfox
Powertoys
Obtainum
Thunderbird
Books
Non Fiction
NNT
Algorithms to Live By
The Pragmatic Programmer
Fiction
Stepping aside from the deluge of technical content above, I want to highlight some of my favorite fiction books/series/authors. You may quickly notice that I read a lot of science fiction and fantasy, alongside the odd classic. THis list certainly suffers from some strong recency bias, and I’m curious as to how it will evolve in a few years.
Terry Pratchett
Basically everything by Pratchett is gold. I think it may be distinctly British humour, but if it clicks for you its exorbitantly funny. I’d suggest starting with Discworld in chronological order, though you could skip a few of the first novels as I’ve gleaned the general sentiment online is that those are worse than the rest. The Guards series is top class and a good place to start–but I really recommend reading the whole series through and through.
Count of Monte Cristo
COMC is probably my favorite single book of all time–if I could pick a series/author Pratchett wins any day of the week but a single work? COMC narrowly edges out some of the Guards/Death/Mort novels to claim my (oh-so-prestigious) number one spot. It is certainly a tome, but it is well worth it. Love, loss, revenge, action, drama, intrigue, mystery, early 19th century French politics–COMC has it all. Just like the Le Trois Musketeers series, these books are adventure novels, high fiction if you will. Read the Robin Buss translation (Penguin Classics.)
Straight Man by Richard Russo
Dr. Chance recommended this, and now I think all (or at least most) department meetings end up with someone clutching their nose after being stabbed by a spiral ring notebook. Very dry, very satirical, very funny. Duck/10; would recommend.
Earthsea
A loose friend of mine recommended Earthsea a number of years ago. She was dead on with her recommendation and Earthsea became my jumping off point to finally reading more by Le Guin. I enjoyed this series, though it did feel a smidge slow or YA-y sometimes–but on a whole would certainly recommend.
Work In Progress
Speaker for the Dead
Dune
Anathem
Wheel of Time
Don’t Let Me Be Lonely
The Ones Who Walk Away from Omelas
Lord of the Rings
Hyperion
Sprawl
Wool
Foundation
Murderbot
Ancillary Justice