Crowbars and cryptographs: my favorite reads from the first half of 2023
This year, I started to actually save down the readings I most enjoyed. It’s made it easier to revisit, relearn, and, by writing about them, re-digest.
This is a long one - jump around! Find the stuff that interests you.
The cabinet of Wikipedian curiosities
Summary: A collection of weird but true pieces of history, as documented on Wikipedia.
Why I liked this: These gems speak for themselves. Some highlights:
The US Air Force wanted to nuke the moon to increase domestic morale, and Carl Sagan was involved.
There are well-documented, unexplained cases of spontaneous dancing in the 14th to 17th centuries.
The CIA’s Acoustic Kitty project went about as well as you’d expect
When the US was on the brink of the Civil War, it fought a mini-war with Canada over a pig
No matter how bad a planner you are, your event will not be as much a shit show as the 1904 Olympic Marathon.
A chicken that lived for 18 months after losing his head.
The immovable ladder of the Church of the Holy Sepulchre
Ming the clam, the oldest living creature ever discovered
The 1922 NYC Straw Hat Riots
The expanding dark forest and generative AI
Summary: The explosion of LLMs will make the Internet an increasingly human-less place. Because of that we will need ways to prove our human-ness online.
Why I liked this: I bookmarked this five months ago, when all the web was aflame with the cooing of the alien baby that is ChatGPT. At the time, the vision of a future internet filled with generic, recycled content - supercharged by LLMs - resonated with me. It still does: the front page of Google is filled with trash and recycled content.
But two things in hindsight:
the OpenAI API has been available for a while now, and at least anecdotally the internet’s ratio of generic content to genuinely creative content seems about the same.
if elements of a network as decentralized as the internet turn royally shitty, We the People will just replace/compensate for that element. This is what a lot of techno-dystopia gets wrong: competition exists, and every piece of damaging technology has a counterpart built to hedge against it.
If Google results become reliably generic and useless, people will turn to something else. I’ve already switched to Reddit as my search engine. Do we really think Google will rule unchallenged for hundreds of years?
If Twitter and TikTok and YouTube get flooded with mediocre AI-generated content and users become less engaged, it will be in those platforms’ interest to adjust their algorithms to reward engaging, unique content.
If AI-generated cold email spam becomes apocalyptic, startups will build better spam filters.
End of the day: it’s fun to try to predict the future of AI, but truly none of us knows how it will go.
Quotes:
You thought the first page of Google was bunk before? You haven't seen Google where SEO optimizer bros pump out billions of perfectly coherent but predictably dull informational articles for every longtail keyword combination under the sun.
We're about to drown in a sea of pedestrian takes. An explosion of noise that will drown out any signal. Goodbye to finding original human insights or authentic connections under that pile of cruft.
Visual design rules you can safely follow every time
Summary: Beginner rules of thumb for visual web design.
Why I liked this: I found this aid useful while I designed app.usebracket.com, my first real design project. Even if you’re not designing a site, it’s worth the reminder how much thought goes into every element you see online. Like item placement at a grocery store, there is a calculated method underlying what we take for granted.
Quotes:
Everything in your design should be deliberate.
Brutalist web design
Summary: The role of a website is to provide enjoyable content to users and to allow them to interact with you. To design and program a site with other priorities in mind is to not prioritize the user.
Why I liked this: At first brush, this is the counterargument to the article above: screw your fancy polished sites, and just give me readable, scrollable text. Up yours, Buffalo City Hall, we’re going with headquarters of the Polish United Workers Party Boston City Hall. There’s something to be said for this approach, and Hacker News is a testament to that.
Ultimately, both of these articles are hitting at the same truth: a website fulfills a purpose, and good design serves that purpose.
Quotes
Brutalist Web Design is honest about what a website is and what it isn't. A website is not a magazine, though it might have magazine-like articles. A website is not an application, although you might use it to purchase products or interact with other people. A website is not a database, although it might be driven by one.
A website is about giving visitors content to enjoy and ways to interact with you.
PCjs Machines
Summary: A PC time machine. Go play OG Space Invaders! Get grateful for G Sheets by trying to use VisiCalc! Die immediately in The Oregon Trail!
Why I liked this: Talk about brutalist design.
The past is a different country. It’s fun to visit that country to remind yourself when it was just a little harder to get addicted to your computer.
A claxonomy of Mexico City’s traffic
Summary: A (beautifully-written) taxonomy of Mexico City’s horn sounds and their meanings.
Why I liked this: In my travel experience, there are two types of countries: those in which people are generally worried about the consequences of doing illegal shit on the road, and those where the people are not. In the former (e.g. the US), we gasp at the audacity of that pickup truck running the red light. Shocked by this violation of our national social contract, we treat our steering wheel as if it’s the criminal’s diaphragm and let loose with a palm heel strike straight from the dojo. When there is no need for communication between cars - when heavy regulation, clear signage, and the ever-present fear of a ticket keep things moving - the horn quickly becomes solely a “fuck you” tool.
But in countries where road laws are, in Captain Barbossa’s famous words, “more like guidelines,” there exists a level of uncertainty that can only be mediated by driver communication. From this, a gorgeous bloom of horn variations, subtle in their differences yet consistent in their meanings. To walk next to a busy avenue in Mexico City is to hear bizarro bird song, a world of communication meaningless to the untrained ear, yet fundamental to the birds’ survival.
Quotes
In the spirit of classical taxonomy, this essay arbitrarily selects a series of common honks to assert an overarching system of meaning shared by people on Mexico City’s streets. While it might sound cacophonous, that residents can distinguish the meaning of each horn shows we’re far from Babel; motorists’ improvisations are a vocabulary emergent from the demands made by a megacity that is, in Dean Chahim’s (2022) memorable phrasing, “governed beyond capacity”. As residents loudly fill the void left by the state with new apparatuses of meaning and management, convention replaces rule so people can keep moving.
As these sounds weave between congested cars, drowning out and being drowned out by other horns, merging into a single force, traveling out of the city to the mountains that hem it, ricocheting back and radiating out anew, they articulate an overarching feature of the city’s social life, one whose restless audibility defies the stasis of what can be seen.
Here’s what it would take to stop or slow AI
Summary: Written in March 2023, Stokes lays out what it would take to actually slow down AI development, which some high-profile folks have called for. Essentially, you can either block consumer/private company access to GPUs, treat model files as contraband, or both. Also, to really slow AI, you’d need to do this with global cooperation.
Why I liked this: “Wars have been primarily won or lost by logistics,” and those who want to slow AI progress face an insane uphill logistical battle. Again, I can’t pretend to know what will happen (nobody can!!), but a globally-coordinated all-out ban on AI development seems less likely than a Butlerian Jihad. The trigger of either would need to be an event so catastrophic and threatening to humanity’s survival that it supersedes national borders, and I imagine our fight-or-flight response will prefer destruction, rather than regulation, of the machines.
In short, in my view, there’s no stopping AI until/unless things get truly bad.
Quotes
Unlike most of the other grasping, clout-chasing hacks who are de facto advocating to bring tyranny and relative poverty to our shores with their calls for constitution-busting restrictions on who can own what hardware and have what files, Yudkowsky and many of his rationalist and Effective Altruist fellow travelers are up-front about the fact that the only thing that can stop AI is an ultra-powerful one-world government with a stunning amount of control over what 8 billion people can and cannot do with all the computers we’ve made.
Saturday morning breakfast cereal: every seven years
Summary: If it takes 7 years to master something, you could treat your life as multiple 7-year opportunities to become great at something.
Why I liked this: Most young people assume that their life will follow a semi-linear path. School leads to maybe grad school leads to a decades-long career of higher pay and greater responsibility.
But what if you put an expiration date on the thing you’re currently doing? What if you told yourself that in 7 years you will die and be reborn, and at that point will need to pick something new?
Sure, this seems idealistic at first brush, and yes the march of time / genetics / bad luck will prevent you from becoming truly great at certain things. But isn’t it more energizing to orient yourself within a framework like this? Isn’t it better to clearly define your era for yourself? Isn’t it better to get 8-10 intentional shots on goal over your life than to let life simply happen to you?
In short, take “deadlines as a technology” (mentioned below) and apply it to your life.
Relevant honorable mention: How to do great work
Juice
Summary: A deep dive into the idea of “Juice,” the non-essential touches and effects that make game or software feel alive.
Why I liked this: As a gamer, Juice is something I’ve always recognized but never formally acknowledged. You can have a site with a decent UI/UX without Juice, but you will not have a site that people love using. You can create a compelling game without Juice, but you won’t form a cult following like Half-Life or Portal.
Juice goes beyond design; it is hospitality. It’s caring for the user’s emotional needs, delighting them without overwhelming them. It is cultivating, as discussed in the next article below, a world that responds to the user and thus feels alive.
Juice is increasingly rare. Juice is not 80/20. In startups, it’s a bad idea to focus on Juice until you have product-market fit. Even after product-market fit, your main problems are probably scalability and reliability. You do not need to prioritize Juice, so most teams don’t.
Deprioritized Juice = the entire internet looking and feeling the same (see the Clif Bar example in the article).
Juice has lessons to teach us beyond gaming and software.
Quotes:
Juice is the non-essential visual, audio and haptic effects that enhance the player's experience.
Something that abundantly fulfills emotional requirements can be described as having soul.
Find what makes you feel, ask why and use that to shape your work.
The Cabal: Valve’s design process for creating Half-Life
Summary: A deep-dive into the creative process behind Half-Life (1998), the debut game from Valve. Valve would go on to fundamentally change the gaming world, both with the titles it produced (Half-Life, Portal, Team Fortress, and Counter-Strike) and with the distribution model it pioneered (Steam).
In this article, we witness how Valve scrapped its original game design, honed in on fun-maximization, and managed to successfully design by committee (the “Cabal”).
Why I liked this: Valve games are masterpieces. Two decades after the first Half-Life games, gamers are still swooning over them, even remaking them with Valve’s explicit consent.
What is Juice? Juice is hitting a vending machine with a crowbar and watching a cola drop out. After playing a Valve game, it’s hard to play anything else.
It’s good to understand the design process that led to these hits. People are wary of “design by committee” approaches, yet Valve managed to pull it off.
My main takeaway, though - perhaps my most important learning this year - is the idea that if the world doesn’t respond to the player, the player will stop caring about the world. How many things can be explained by this? How many dissatisfied employees, unhappy spouses, depressed kids? How many players are hitting the vending machine with a crowbar to no effect? And whose job is it to fix that?
Quotes:
The first theory we came up with was the theory of "experiential density" — the amount of "things" that happen to and are done by the player per unit of time and area of a map. Our goal was that, once active, the player never had to wait too long before the next stimulus, be it monster, special effect, plot point, action sequence, and so on.
The second theory we came up with is the theory of player acknowledgment. This means that the game world must acknowledge players every time they perform an action. For example, if they shoot their gun, the world needs to acknowledge it with something more permanent than just a sound — there should be some visual evidence that they’ve just fired their gun … Our basic theory was that if the world ignores the player, the player won’t care about the world.
A final theory was that the players should always blame themselves for failure. If the game kills them off with no warning, then players blame the game and start to dislike it. But if the game hints that danger is imminent, show players a way out and they die anyway, then they’ll consider it a failure on their part; they’ve let the game down and they need to try a little harder.
In order for highly hierarchical organizations to be effective, they require one person who understands everyone else’s work at least as well as the individuals doing the work, and other people who are willing to be subordinates yet are still good enough to actually implement the design. Given the complexity of most top game titles, this just isn’t practical — if you were good enough to do the job, why would you want to be a flunky? On the other hand, completely unstructured organizations suffer from lack of information and control — if everyone just does their own thing, the odds that it’ll all fit together in the end are somewhere around zero.
Ron Cobb’s semiotic standards for Alien (1978)
Summary: This one’s quick, screenshot is below :)
Why I liked this: Semiotics is the science of meaningful symbols. Before computers laundry symbols were maybe the most commonly-seen (and least-understood) set of semiotic standards.
If you had to design the equivalent of laundry symbols for a space fleet, how would you design them? How would you make them simple, memorable, and meaningful? This is Ron Cobb’s answer, and you can see them in Alien.
Relatedly, one of my favorite species-wide questions is “How do we convey the danger of our nuclear waste to far-future humans?” Since we can’t assume they speak our language, or even recognize the same symbols, this is a unique semiotic challenge.
Stories we’ve seen too often
Summary: A list of plot types that the editors at Strange Horizons, a speculative fiction magazine, see too often among submissions. These plots are not necessarily bad, they’re just commonly submitted.
Why I liked this: It was humbling to read through this and realize how many “genius” short story ideas I’ve come up with over the years fall into one or another of these :)
Quotes (ie plots):
Weird things happen, but it turns out they're not real.
In the end, it turns out it was all a dream.
In the end, it turns out it was all in virtual reality.
In the end, it turns out the protagonist is insane.
In the end, it turns out the protagonist is writing a novel and the events we've seen are part of the novel.
Technology and/or modern life turn out to be soulless.
Someone calls technical support; wacky hijinx ensue.
Someone calls technical support for a magical item.
Someone calls technical support for a piece of advanced technology.
The title of the story is 1-800-SOMETHING-CUTE.
White protagonist is given wise and mystical advice by Holy Simple Native Folk.
Protagonist agrees to go along with a plan or action despite not having enough information about it, and despite their worries that the thing will be bad. Then the thing turns out to be bad after all.
Deadlines as technology
Summary: No productivity tool in the world will work as effectively - or as cheaply - as good old fashioned deadlines.
Why I liked this: One of the hardest things about running a pre-product market fit startup is constantly setting deadlines for yourself. You have 1) one real deadline (running out of money), 2) some scary yet vague deadlines (you or a cofounder losing hope), and then 3) a bunch of intermediate deadlines that you must define and hit to avoid (1) and (2).
YC instills deadlines-praxis during its batches by forcing you to set ambitious two-week goals, then asking you to explain why you did not hit those goals in the (common) event that you fail.
Quotes:
The only technology that you need is deadlines.
Set a deadline, even if the one thing you learn when you fail to ship, is that you’re bad at estimating and meeting deadlines. That’s a chance to improve for the next one you set.
And best of all? There’s no yearly subscription. Deadlines are free.
Non-Ordinary States of Consciousness Contest: Psychedelic Cryptography
Summary: Can you design a hidden message that only people tripping on shrooms can decode? That’s the question behind this research competition.
As it turns out, yes, psychedelic cryptography is real.
Why I liked this: This is not some hippie-dippie “dude, what is real meaning anyway” college dorm experiment. The visual designers used tracing effects, which are heightened to those on psychedelics, to get their messages across to the tripping judges. And in at least three cases, it worked! Those groovy short videos convey something that we, the sober, cannot grok.
Imagine running with this and designing a short film that reveals layers of meaning depending on how much psilocybin is running around your system. Talk about Straussian, man.
Editor’s note: no, I haven’t tried this myself. Not yet.
Read something wonderful
I leave you with this: a collection of history’s great essays and speeches, from Jonathan Swift to Joan Didion to DFW to Paul Graham. This site is my favorite project of 2023.
Happy reading, friends.