Last time I completely forgot to mention my other fun holidays project. This one’s even sillier than the emojigarden.
I’ve been an admirer of what I consider to be The Art of Funny Tweets for many years now and have amassed quite a collection of Screenshots of Funny Tweets. Two weeks ago, without a reason I can articulate, I decided I needed to share them.
But instead of just putting up a page with a bunch of screenshots on them, Claude and I conspired to create a Python script that syncs screenshots from a Google Drive and then renders them, at build time, into a Markdown page. Claude did the heavy lifting, I only put screenshots in a folder. That makes it very easy for me to add new ones, which is, of course, important for this very important project.
What we ended up with is this page: Funny Tweets. Hope you get a laugh out of them.
Let’s kick this off with a hot take: Go is a Well-Designed Language, Actually — I agree 100% and I think the people who say Go is not well-designed should instead say that they don’t agree with its design goals or that they don’t know what tradeoffs are.
Matt Keeter’s Fidget is “a library for representing, compiling, and evaluating large-scale math expressions, i.e. hundreds or thousands of arithmetic clauses” and that’s one of the few things I understood on that page. After scrolling through it, I wish I understood it all.
David Crawshaw, co-founder of the fantastic Tailscale, on how he programs with LLMs. Insightful, calm, very good.
Goldmine: Applied LLMs. Single page, bursting with valuable information. 5 out of 5.
Test Driven Writing, using LLMs. Fascinating. I’ve done ~TDD with LLMs before and I can see this possibly being helpful, but at the same time I still haven’t warmed to the idea of using LLMs when writing prose and probably never will.
Sean Goedecke writes about “what makes strong engineers strong” and says: “Some of the smartest people I’ve worked with - in terms of raw brainpower - were not particularly effective engineers, because they struggled with pragmatism and speed.” I agree 100%. I also agree with this: “To be a strong engineer, you need the raw confidence to believe that you will figure it out, whatever it is.” I wrote about the same idea last year and still think it’s one of the most important things I need to work on.
Sean Goedecke again, this time on his engineering values. Great page. It’s honest. “I’m more excited by maintenance than rewrites.”
Talking about smart people and being smart: Dwarkesh Patel interviewed Tyler Cowen. I haven’t listened to the whole thing yet, but the first 30 minutes have been very interesting. I loved this bit about IQ: “If you look at labor market data, the returns to IQ as it translates into wages, they're amazingly low. They're pretty insignificant. People who are very successful, they're very smart, but they're people who have, say, eight or nine areas where they're like, on a scale of 1 to 10, they are a nine. Like they have one area where they're just like an 11.5 on a scale of 1 to 10. And then on everything else, they're an eight to a nine and have a lot of determination. And that's what leads to incredible success. And IQ is one of those things, but it's not actually that important. It's the bundle, and the bundles are scarce. And then the bundles interacting with the rest of the world.”
On Linear: “it's very very clear that Linear's success as a product is simple: it is really, really good, in a way that is almost uninteresting. Linear offers very few novel features and instead invested a lot of time, energy, and polish in... just getting everything correct, and making it work extremely extremely fast.” Never used Linear, now I want to.
An operating system in 1000 lines. I’m a big, big fan of “X in <Y lines” projects and this looks very cool. “The tricky part of creating your own OS is debugging. You can't do printf debugging until you implement it.“
I’m actually not sure whether I should share it, but… okay, yes, here, go try out the simulation clicker. Then, after ten minutes, after you thought “what’s wrong with me”, you come back here.
Steve Krouse, from val.town, sharing what they “learned copying all the best code assistants”. I’ve worked on two “assistants” in the last two years and find the whole space incredibly fascinating at the moment. There’s so much happening: inference speed is only going up, new techniques are constantly introduced (speculative decoding!), there’s a lot of twists and remixes of similar ideas, and Steve gives a really good overview here.
For over a year I’ve been searching for this quote, now I finally found it again. Jerry Seinfeld on Howard Stern: “Your blessing in life is when you find the torture you're comfortable with.”
Very cool idea: “Claude projects are truly incredible. My favorite is a version of ‘morning pages.”’I could never sit and write for 30 min so instead I do it with Claude as an audio note and have it organize it.”
I don’t know how or why but the idea of agency showed up in quite a few, quite different conversations over the last few weeks. Agency — what a word, what a concept. I value it dearly, I know it when I see it, but it’s hard to pin down what exactly it is and I often reach for examples or analogies. Last week I got lucky and someone shared a new way to think about agency with me. He said: “Imagine you’re travelling through Middle America and get kidnapped. Someone mistakes you for someone else. A mixup. You end up in a prison in the middle of the jungle. You’re allowed one phone call. One phone call to a person to get you out of there. Who do you call? The number one person on your list, that’s a high-agency person.”
In another conversation I mentioned Simon Sarris’ excellent The Most Precious Resource is Agency, which has been updated to School Is Not Enough and which I constantly think of and refer to, always with “Look, I’m not saying we should bring back child labour, but…”Television, a “blazing fast general purpose fuzzy finder TUI”, looks like a really nice homage to TJ’s telescope.nvim, which I love.
David Foster Wallace on complicated books: “And I hate books where, you know, those books where you get halfway through and you get the sense that the author is so stupid that he thinks he can fool you into thinking that the book is really sophisticated and profound just because it's difficult. It's an epidemic in academic writing. And it happens about half the time in avant-garde writing.”
There’s a lot more interesting stuff in here than the Buzzfeed-y title might suggest: The 30 Best Pieces of Company Building Advice We Heard in 2024. I love the bit about the “set of 12 questions that Stripe leaders ask in every product review” with “a very strict guidance that you cannot move on to question two until you've sufficiently answered question one”. Question one? “Who is the target user?” Now think of all the product development you’ve witnessed in your life — was there an answer to that question? (I also love #16 on reference calls.)
Big pirating nostalgia when reading this New Yorker piece on Spotify: Napster, LimeWire, Kazaa get a mention. I’m missing eDonkey2000, eMule, and Morpheus though. It’s an interesting piece, but extending the historical view to the past by, say, 60 years, would’ve made it not just more balanced but also more interesting — musicians earning a lot of money with their music is — relatively speaking — a recent phenonemon.
