🌀🗞 The FLUX Review, Ep. 150
June 13th, 2024
Episode 150 — June 13th, 2024 — Available at read.fluxcollective.org/p/150
Contributors to this issue: Dart Lindsley, Ben Mathes, Erika Rice Scherpelz, Dimitri Glazkov, Neel Mehta, Boris Smus, MK
Additional insights from: Ade Oshineye, Justin Quimby, Alex Komoroske, Robinson Eaton, Spencer Pitman, Julka Almquist, Scott Schaffter, Lisie Lillianfeld, Samuel Arbesman, Jon Lebensold, Melanie Kahl, Kamran Hakiman, Chris Butler
We’re a ragtag band of systems thinkers who have been dedicating our early mornings to finding new lenses to help you make sense of the complex world we live in. This newsletter is a collection of patterns we’ve noticed in recent weeks.
“He who knows all the answers has not been asked all the questions.”
— Confucius
📝 Editor’s note: We can’t believe it’s been 150 episodes and just over three years since we launched FLUX! Thank you for your continued support!
🦉🔮 Preserving the future
A team has finally received the mandate they’ve been seeking: a complete overhaul of a core part of their product. Although they need to maintain some level of backward compatibility and ensure no data is lost or corrupted, they have broad authority to change both the product experience and its technical underpinnings. They’re excited and ready to go!
However, the one person who has been on the team for years, who knows the system inside and out and where the skeletons are buried, is wearing a skeptical look. They are the ones who always seem to speak up and point out the risks. Sometimes seen as a curmudgeon, they’ve also saved the team from enough mistakes that it’s worth taking this skepticism seriously.
This paradox lies at the heart of system ownership and maintenance. Whether it’s a classic like James C. Scott’s Seeing Like a State or Joel Spolsky’s essay “Things You Should Never Do, Part I” or a newer entrant like Ashley Goodall’s The Problem with Change, we often get advice that there’s wisdom in the past. This same wisdom is captured in Chesterton’s fence, which advises us not to tear down a fence whose purpose we do not know.
Yet, as architects of systems, be they software, social, physical, or otherwise, we are responsible for shaping their future. The very idea of improvement is founded on the notion that the past has not given us what we need (at least, not what we currently need). We sometimes find ourselves in the awkward position of agreeing with both perspectives: preserve and improve.
Navigating this paradox requires taking a deeper look at the notion of preserving the past. In a world sometimes addicted to change, it asks us to assume that people in the past were at least as thoughtful about making decisions in their context as we are now, and that some of the forces impacting them are likely still impacting us. Chesterton’s fence doesn’t say “never touch the fence.” It says, “only change the fence once you know why it was put there.”
Both dismissal and veneration of the past get us into trouble. By dismissing the past, we miss out on its lessons — the point often made by advocates for preservation. On the other hand, if we venerate it, we end up stuck, maintaining structures that no longer deliver value — or which even cause harm in the current context.
Instead, we are better off aiming to understand the current system. What decisions were made over time? What assumptions went into those decisions? Which of those are still true? Which have changed? Which parts of the system are still helpful? Even if we don’t know the actual history of a system, we can explore hypotheses. We’ll know we’ve gone deep enough when we see how the system’s current state makes sense for ourselves.
Once we understand the system, we gain agency. If we have the power to change the system, we can do so, preserving the good while updating that which no longer makes sense. If we don’t have the power to change the system, we can at least understand just how much to play the game to avoid getting too injured by the system’s flaws.
Of course, no one can ever understand a complex system perfectly. At best, we can get a fuzzy picture of the whole and, perhaps, a clear picture of some parts. However, striving for understanding can help us better navigate the paradox of preserving and improving systems. This balanced approach allows us to create resilient and adaptable systems ready to meet current needs while honoring the wisdom of the past.
🛣️🚩 Signposts
Clues that point to where our changing world might lead us.
🚏🔌 Installing new power lines could double the capacity of the US’s energy grid
One of the biggest factors slowing the growth of renewable energy is that the US’s grid simply can’t handle all the power energy producers want to push through it. The country has also been building power lines extremely slowly, with the average line taking 10 years to complete. But experts have found a quick, cheap, and huge win: simply installing new carbon-fiber wires would double the amount of power that could flow through the grid and provide up to 80% of the new transmission capacity needed. And since no new wires or towers need to be built, it’s a cheap upgrade; it’s also fast to implement since no new right-of-way has to be acquired.
🚏☎️ “Emotion canceling” AI can make customers sound calmer to call center reps
Japanese giant SoftBank announced that it’s created voice-altering technology to soften people’s tone and voice while keeping the words the same. Softbank says the technology could help reduce customer support agents' stress when dealing with angry customers all day. If a customer conversation gets too long or “too abusive,” the software will warn the customer that it might hang up.
🚏👹 The new Stable Diffusion model is producing horribly grotesque pics of humans
Stable Diffusion 3, the latest version of the popular image generation model, has drawn criticism for producing bizarre, mangled pictures of humans with upside-down heads, backward hands, arms sticking out of their backs, or bodies made only of legs. Commenters called it a notable step backward from previous versions of SD, and some theorized that Stable Diffusion’s zeal to scrub out pornographic pictures might have left the model without enough training data to understand what humans look like.
🚏💊 Americans are paying overseas workers to help them find Adderall
The US has been suffering from a well-known shortage of Adderall, and weight-loss drugs like Ozempic are in short supply as well. So, for $50, one “concierge” service will have a “virtual assistant” abroad — often in the Philippines — do the time-consuming work of calling all the pharmacies in your area to see if they have some of these hard-to-find medications in stock. One happy customer reported that they’d been looking for one particular medication for two months, yet their assistant found it in just two days.
📖⏳ Worth your time
Some especially insightful pieces we’ve read, watched, and listened to recently.
Browser Extensions Are Underrated: The Promise of Hackable Software (Geoffrey Litt) — Argues that browser extensions are the rare exception in a world of closed platforms that let users modify the apps they use.
How Floating Point Works (jan Misali) — Starting from first principles, builds up an understanding of how computer architects designed the floating-point standard for storing large numbers and decimals. As one commenter says, it’s a great example of the “pretend that you’re inventing it for the first time” teaching method.
The Struggle for the Soul of Milton Keynes (The Guardian) — Describes the idealistic and surprisingly forward-looking vision behind the UK’s oft-maligned planned city. MK’s emphasis on a modular grid, open common spaces, and pedestrian walkways separated from roads is unique, and longtime residents are locked in a battle with developers who they say don’t appreciate the unique character of the city.
The Beauty of America’s Ugliest Ballpark (New York Times) — A paean to the gritty, freewheeling, and affordable Oakland Coliseum, which will see its last baseball game at the end of this season when the Athletics decamp for Sacramento. In a league where every team has settled on a polished retro-modern ballpark with corporate boxes and plush amenities, there’s something special about “baseball’s last dive bar” — a phrase that this essay made famous.
🔍🔺 Lens of the week
Introducing new ways to see the world and new tools to add to your mental arsenal.
This week’s lens: Trilemma.
You and your project colleague are puzzling over a problem: the product you build is laser-focused on its users and what’s good for business, but there’s still insufficient lift to hit your goals. It’s almost like there is some other unseen force tugging at you. After a lot of whiteboarding, exhausted, it dawns on you: you forgot about the surrounding ecosystem. There aren’t two parts in this puzzle, but rather three. And you can only have your product focus on two of them at a time.
Welcome to the world of trilemmas. Trilemmas are like dilemmas, but instead of two unfavorable or mutually exclusive choices, there are three. Trilemmas typically represent impossible spaces: we want to satisfy all three conditions somehow but can only afford their subset.
Trilemmas are all around us. The infamous project management triangle is a trilemma: among scope, cost, and time, you can only pick two. There’s also Zooko’s triangle, which posits that a network protocol can only satisfy two of the desired constraints (memorable, secure, or decentralized).
Unlike some lenses, trilemmas don’t offer solutions or paths forward. Instead, they clarify the situation in a way that allows the participants in this situation to recognize and respect the impossibility of its full resolution.
Recognizing that the scope of a project can’t keep growing without impacting cost or time is a very valuable insight — and something that a skilled project manager will know by heart.
A good way to uncover a trilemma is to listen to an argument or discussion going in circles and try to discern a missing variable. If all assumptions in the conversation are implicitly based on different assumptions about the values of this variable, you might have a trilemma on your hands.
© 2024 The FLUX Collective. All rights reserved. Questions? Contact flux-collective@googlegroups.com.