Link Log

The Cybertruck (aka The World's First Ecofacist Vehicle) is dead, after at least seven recalls and who knows how many driver, passenger, and bystander casualties.

I will mark this occassion by relating a conversation I overheard while walking through City Park in New Orleans:

Friend 1, squinting at a parked Cybertruck: So...it's like a truck, but it's a Tesla?

Friend 2: Yep, that's pretty much it.

Friend 1: ...well, why does it look like that? It looks like shit.

“Oh, neat!” is your get out of Hell faster phrase.

My version of this is "Good", preferrably said in your gruffest, toughest voice. I picked that one up from Struthless. "Oh, neat!" is pretty good too, because it puts you in the mode of "I'm going to treat this problem like a cool bug that I'm studying with my magnifying glass."

The article outlines ten ways that "Oh, neat!" works for you.

Mundango is your daily bingo card of mundane joys and I love it.

Designer Bruce Mau lays out some of his techniques for "growth." The techniques are valuable, and I'll probably borrow a few. But what I found most enlightening here is that Bruce's implicit definition of growth is a few degrees off of how the dominant culture defines it. Growth here isn't business growth, and it isn't even "personal growth" defined through the lens of productivity. I think when Bruce says "growth," he purely means "making better stuff" and I think that's the most powerful thing.

A presentation from Phillip Rogaway from NIST's Workshop on Block Cipher Modes of Operation. It was maybe not what the audience was expecting, and that is the point.

What is the small thing we've been lied to about? The "Standard Technological Narrative." My problem with the STN: It's a fantasy

Phillip gives his suggestions for building a Radical CS to combat the STN, and walks through some of his implementations in papers and teaching. More of this please!

NESFab is a new programming language for creating NES games. Designed with 8-bit limitations in mind, the language is more ergonomic to use than C, while also producing faster assembly code. It's easy to get started with, and has a useful set of libraries for making your first — or hundredth — NES game.

I love a good single-purpose language. Back in my early computing days, special purpose compilers with their own dialect and a single target were all over the place. Today, the vast majority of development is done in "write once, run everywhere languages." Obviously, if you want to reach the most users and platforms, there are advantages to that.

But if all you want to do is run on one device? There's power in specialty tools.

A TTRPG in the style of Lasers & Feelings, but:

YOU ARE THE EMPLOYEES OF THE STARTUP RAPPTR. Your job is to scale the company, interface with customers both friendly and deadly, and defend fellow employees against workplace dangers. CEO CHAD has just come back from Burning Man and is overcome by a strange psychedelic energy, leaving you to fend for yourselves while he recovers in a napping pod.

Your character has two stats: Hard Code (science, reason), and Soft Skills (rapport, passion).

A good reminder. I've seen a lot of folks decrying how it is possible that "so many people" voted for Trump. The truth is, as of election night, he had a tiny margin in the popular vote. In his other two elections, he lost by a larger margin. This is, of course, after decades of voter suppression by the GOP, so it's hard to say he's genuinely favored by the majority of people in this country.

Tuki is an open source utility for adding a little structure to running one-off commands in production. It's an interesting approach to a problem that I've encountered at a lot of clients. I'm actually working on a slightly different solution (and obviously I like my version more or I wouldn't be building it!).

Some of the people behind Disco Elysium have formed a new co-op to make new games and their manifesto rules. You can tell this was written by BAFTA-winning writers.

Joseph gives a quick overview of Anom, the encrypted phone company beloved by high-end criminals around the world. The wrinkle: Anom was secretly being run by the FBI.

I've added Dark Wire, Joseph's much more in-depth book on Anom to my pile for reading soon!

A variety of delightful videos that line up really well with songs.

I'm prone to "Owl-speak." As usual, it's a good idea to be more like Pooh Bear.

There is no such thing as multi-tasking. Even computers don't run more than one task at a time (at least on a single CPU core / "brain"). What they do is time-slice. But time slicing has a cost if you do it too often, especially if you aren't a computer.

A classic of programming lightning talks.

Well worth the read if you make software on an "agile" team. First, Barry explains why "story points" don't work. Then Barry outlines what to do instead.

It's brilliant to apply queueing theory and Little's Law to estimate delivery of software in a queue. As software engineers we know it works for job and message queues. Why wouldn't it work for development queues?

A great take on the concept of "technical debt."

Two key takeaways:

1. "Slow is smooth, smooth is fast."
2. It's only technical debt if it's intentional. Otherwise, it's just bad code.

It's good to be reminded sometimes.

I saw this guide linked from the Elixir Slack and it is the best guide to comprehensions and the for special form in Elixir that I've seen. I write Elixir code every day and I had no idea about some of the options. See also: the official documentation for comprehensions.

This is a great search engine for public domain images.

Apple's implementation of RCS is a perfunctory, letter-of-the-law effort. Their messaging blames the EU instead of acknowledging their half-assed approach toward the technology. And prominent Apple bloggers are providing cover for it, instead of using their influence to hold the company to account.

This is an example of what a software update that actually serves users looks like-- something that is sadly a rarity these days.

Car companies: Dare to do less · I couldn’t possibly walk away from our time in the Ranger without thinking about the absolutely insane amounts of money and resources and carbon loading we could save by building smaller, simpler, cheaper, dumber, automobiles.

Please, yes. In the last 15-20 years, auto makers, particularly US auto makers have completely abandoned the basic sedan market and the actual light truck market. Before that, base sedans and trucks were intentionally sabotaged (e.g. base trim without aircon) so that dealerships could upsell you expensive packages (e.g. aircon plus a bunch of shiny bits you don't care about for $5,000).

I feel like there's an opportunity for someone to change the game with a car like this:

• ✅ A stero system that is basically a dash-mounted $20 bluetooth speaker and one USB port
• ✅ Aircon.
• ✅ An eye toward user-serviceable.
• ✅ Cheap, reliable, and durable.
• ❌ No other computer-y or electronic bits.

I love so much about this presentation. I believe in the power of writing simple, rough software for small communities, or even for communities of one (e.g. yourself). I've done a lot of that. I also have been feeling the pull of local-first software with really simple tech stacks. That might be my disillusionment with the complexity of modern DevOps.

That said, I don't buy that LLMs are going to suddenly allow a lot of people who have never coded before to start coding. Every LLM-based coding tool I've tried has been basically IntelliSense, but with more variance (plus and minus). To say that people without any other assistance can use IntelliSense to develop whole applications would be nonsense, and so far, I think it's the same for LLMs. Now, does richer developer tooling lower the barrier to entry? Certainly.

My system is pretty similar to this, though I'm still living the one-file-per-day life. I've tried just about all the options out there, and for me, the one that has the lowest friction and highest interoperability is text files.

It also reminds me of the adage I've heard repeated many times about David Allen's Getting Things Done system: "if you can't do GTD with a pen and paper, you aren't really doing GTD."

As an old punk, I love this.

Though, I'd much rather be in a terrible mosh pit than an average social media site.

Cool idea— a font to replace distracting lorem ipsum.

I feel like this solution isn't ideal. My preference is for writing-first design i.e. forcing the client to write their message before you start creating high-fidelity design assets. And I think if I were in a situation where placeholders were appropriate, I think I would prefer something that still felt like glyphs. I feel like the first question from a client would be "what's with all of these blocks everywhere?"

If you don't know Struthless (Campbell Walker) yet, he's making some of the best video essays about creativity, life improvement, mindset, and internet culture on YouTube right now. Go check out his stuff.

In this one, he tackles the "radioactive" subject of masculinity. I know what you're thinking. Uh oh. Is he for or against masculinity? Which camp is he in? Watch the video.

Yeah, this pretty much sums it up.

Neil Stephenson wrote The Diamond Age: Or, A Young Lady's Illustrated Primer, a novel featuring an AI storybook which teaches the protagonist. A Redditor asked him "My ultimate goal in life is to make the Primer real. Anything you want to make sure I get right?." Neal simply replied "Kids need to get answers from humans who love them."

I wish the generative AI people would take that wisdom to heart.

Watching in real time as “slop” becomes a term of art. the way that “spam” became the term for unwanted emails, “slop” is going in the dictionary as the term for unwanted AI generated content.

Slop. I love (/hate) it.

Essentially, Scrum Masters are corporate dog trainers. Command, control, and structure strengthen the pack and the bottom line. Once they're trained, it's time to move on

I wish more people in the software business understood this. Having permanent, full-time Scrum Masters is an anti-pattern. Their prescence, at least in the long term, is a sign that you don't actually have an empowered agile product team.

The thing is: none of this is gone. Nothing about the web has changed that prevents us from going back. If anything, it's become a lot easier. We can return. Better, yet: we can restore the things we loved about the old web while incorporating the wonderful things that have emerged since, developing even better things as we go forward, and leaving behind some things from the early web days we all too often forget when we put on our rose-colored glasses.

The fascinating origins of Oneida silverware.

I think the devil is real and he wants you to be more productive. He’s everywhere, spreading wickedness disguised as wisdom.

...

For legal reasons, I’m not saying the people who write this stuff are literally Lucifer in human skin. It’s just that, if I wanted to maximize human misery, I would 100% try to convince people to spend more time doing things they hate.

The bezzle is a term originally coined by John Kenneth Galbraith for a long-term pattern of bad faith in which the mark does not realise at the time that they have been a victim, and may even feel that they have gained in the short term, until being disillusioned later on. The term is a contraction of the word "embezzlement". The bezzle does not necessarily require criminal acts; the creation of illusionary wealth suffices. I think we're living in the bezzle.

In the hours following the release of CVE-2024-1086 for the project The Linux kernel, site reliability workers and systems administrators scrambled to desperately rebuild and patch all their systems to fix a vulnerability that allows an attacker with unprivileged command execution to gain read/write access to page tables. This is due to the affected components being written in C, the only programming language where these vulnerabilities regularly happen. "This was a terrible tragedy, but sometimes these things just happen and there's nothing anyone can do to stop them," said programmer King Bud Hodkiewicz, echoing statements expressed by hundreds of thousands of programmers who use the only language where 90% of the world's memory safety vulnerabilities have occurred in the last 50 years, and whose projects are 20 times more likely to have security vulnerabilities.

The magical claim of machine learning is that if you give the computer data, the computer will work out the relations in the data all by itself. Amazing!

In practice, everything in machine learning is incredibly hand-tweaked. Before AI can find patterns in data, all that data has to be tagged, and output that might embarrass the company needs to be filtered.

...

Commercial AI runs on underpaid workers in English-speaking countries in Africa creating new training data and better responses to queries. It’s a painstaking and laborious process that doesn’t get talked about nearly enough.

AI doesn’t remove human effort. It just makes it much more alienated.

It appears that Shel Silverstein predicted ChatGPT all the way back in 1981.

I can’t stress this point enough. The reason why GAMM and all its little digirati minions on social media are pushing things like crypto, then the blockchain, and now virtual reality and artificial intelligence is because those technologies require a metric fuckton of computing power to operate. That fact may be devastating for the earth, indeed it is for our mental health, but it’s wonderful news for the four storefronts selling all the juice.

[Gunpei Yokoi, long-time Nintendo designer and producter,] said "The Nintendo way of adapting technology is not to look for the state of the art but to utilize mature technology that can be mass-produced cheaply." He articulated his philosophy of "Lateral Thinking of Withered Technology" (枯れた技術の水平思考, "Kareta Gijutsu no Suihei Shikō") (also translated as "Lateral Thinking with Seasoned Technology"), in the book Yokoi Gunpei Game House. "Withered technology" in this context refers to a mature technology which is cheap and well understood. "Lateral thinking" refers to finding radical new ways of using such technology. Yokoi held that toys and games do not necessarily require cutting-edge technology; novel and fun gameplay are more important. In the interview, he suggested that expensive cutting-edge technology can get in the way of developing a new product.

ht Simon Willison.

Everyone has a list of products they've seen fail and disappear or get acquired and disappear. I want a less-ambitious, lower-scale, bootstrapped company with a plan. I love solo founders with low costs and a fair subscription price.

Yep. What deranged times we live in that it is routine not only for businesses to launch with no plan for ever generating a sustainable income, but for those businesses to rake in buckets and buckets of investment.

A harrowing tale of fixing an unsustainable software project. I've been there.

After being there, I agree with one of Luka's takeaways-- you cannot wait for permission to fix maintainability problems in software. You have do it. And if you have to do it quietly and subvert "the process" to make it happen, do it anyway.

I love to hear these kind of stories about adopting Elixir. I went through a similar arc when I moved from Python/Django, to Node, to Ruby/Rails, and finally to Phoenix/Elixir.

There's a swing back to RSS right now, which I think is good. But I also think that Evan's thoughts here are good. RSS can't be the only solution for how we take in the web. It can't be the only solution for how we decentralize the web again.

Hat tip to Greg Morris for his related post that helped me find this one.

A collection of quotes about AI from 1683 to 2024.

A level-headed explanation of what exactly AI, ML, and LLMs are. "LLMs are really complex markov chains; but the really complex part makes them qualitatively different" has been by go-to explanation of LLMs. I disagree, however, with the idea that LLMs "understand" anything. LLMs contain big statistical models of sentence and paragraph structure, and of the relationship between words and phrases. This allows them to generate text that is more of a statistical match for text written by humans. That humans see this as "understanding" is a form of pareidolia. This distinction is narrow, but important.

A series of great principles for a better web! Hear, hear! Side note: thanks to Cory for inspiring me to add a link log feed to this site, inspired by his at coryd.dev.

Leandro argues that the key to writing readable code is taking implicit context and making it explicit. I agree! This is also one reason I love Elixir. Many of the design choices in the language and standard library encourage you to explicitly write what you mean-- even if it is more keystrokes.

A brilliant lens for thinking about our relationship with tech products, design, and terms of use. Opt-out "consent" isn't consent at all. "Maybe Later" isn't consent either.