🌀🗞 The FLUX Review, Ep. 99
May 11th, 2023
Episode 99 — May 11th, 2023 — Available at read.fluxcollective.org/p/99
Contributors to this issue: Justin Quimby, Neel Mehta, a.r. Routh, Boris Smus, Erika Rice Scherpelz, Jon Lebensold, Dimitri Glazkov, Ade Oshineye
Additional insights from: Gordon Brander, Stefano Mazzocchi, Ben Mathes, Alex Komoroske, Robinson Eaton, Spencer Pitman, Julka Almquist, Scott Schaffter, Lisie Lillianfeld, Samuel Arbesman, Dart Lindsley
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.
“Actually, in human affairs, it is often next to impossible to break things neatly up into ‘inside the system’ and ‘outside the system’; life is composed of so many interlocking and interwoven and often inconsistent ‘systems’ that it may seem simplistic to think of things in those terms.”
— Douglas R. Hofstadter, Gödel, Escher, Bach
🫣🧬 The hidden complexity within ‘us’
The world is made of systems: sets of entities and the tangles of relationships between them. Programs are systems. Computers are systems. Organizations are systems. Ecosystems are systems. We, as individuals, are systems. We are surrounded by systems, from our inner monologues to the stars twinkling in the sky.
When dealing with systems, Ashby’s law of requisite variety is a powerful tool.
Each system has a variety of states. A simple latch might only have two: locked or unlocked. More complex systems demonstrate more variety. Back in the day, we could gauge the complexity of a computer system by counting the transistors, but even the most primitive modern computer has massively more states. Biological systems easily outscale those; with its chemical and physical variations, even a simple cell has an unimaginably immense number of possible states.
When we look to modify a system, we need to understand the states it might have. The law of requisite variety states that for one system to control another system, the controlling system must have a greater variety than the system being controlled. We must contain more complexity than a latch to control the latch. We must be more complex than the relevant aspects of a computer to control it. When we attempt to control a system with a larger state space than we can contain, at best we control a simplified model of the system that works well enough. At worst, we fail. We may find that, if anything, this other system controls us.
Then… how do organizations function? If we are, individually, immensely complex systems, then a gathering of us is even more complex. How are we able to align and organize? There’s a seemingly intractable problem here: to lead an organization, the leader must be at least as complex as the organization, but the leader is also part of the organization, so that’s impossible! No matter how brilliant or effective a leader is, they cannot encompass all of the complexity of an organization they are a part of.
What happens instead is that we elect to simplify ourselves. As part of an organization, we project ourselves onto simpler systems — Engineer, Product Manager, User Experience Researcher, Barista, etc. Each of these is a label which takes a beautifully complex individual and reduces the variety of their states to something more… manageable. Understanding this leads to some important insights.
First, organizations are infinitely more complex than they appear. These simplifying reductions of state only hide the complexity. Turbulent times often peel back the abstractions and showcase this hidden complexity. It is a costly mistake to confuse highly legible roles with the reality of individuals. The notion of “control” in leadership is at best a temporary illusion.
Second, it is humbling and awe-inspiring to recognize the gift of others who voluntarily agree to pretend to be much simpler than their full selves in order to make the organization work. These people who we lead are reducing their variety of states to enable the larger group to function. Good leaders are those who retain this awe and humility and understand the preciousness of the gift of requisite variety — and treat it with appreciation and kindness it deserves.
Finally, we need to avoid the trap of thinking that sacrificing our variety (that is, our individuality) for the benefit of the group is an unalloyed good. As with any sacrifice, this comes at a cost to the individual — suppressing what makes you, you can lead to confusion, frustration, and a feeling of being unmoored or inauthentic. Leaders need to be aware that team members are voluntarily flattening themselves, and as such, they’re free to withdraw if they no longer have confidence in the leader/group or no longer believe that their actions are helping.
The key is to create cultures where we recognize and appreciate people’s self-flattening, use it to drive a collective benefit, and give people space to be their full selves whenever possible.
🛣️🚩 Signposts
Clues that point to where our changing world might lead us.
🚏🗞 China arrested a man for generating fake news with ChatGPT
According to Chinese authorities, a man was arrested for using ChatGPT to write a fake story about a train crash that supposedly killed nine construction workers; the story racked up over 15,000 clicks on social media. This arrest (the country’s first generative-AI-related arrest) drew upon a new law that forbids people from using AI “to produce, release, and fabricate untrue information.”
🚏🏡 Rising mortgage rates are “trapping” people in homes, driving down inventory
Many Americans bought homes or refinanced their mortgages during the low-interest-rate environments of 2020 and 2021. The problem is that many of them can’t sell their homes and move now, because their new houses’ mortgages would come with unaffordably high interest rates. This has had the knock-on effect of reducing the number of homes that come onto the market, thus making it even harder for people to buy. Real estate economists say we’re in a “unique market condition” where “sales are down and even prices are down in some areas,” but it’s still very hard to actually land a house.
🚏🛐 Religious chatbots in India are giving dangerous advice, including condoning violence
One Indian software engineer launched “GitaGPT,” an AI chatbot that plays the role of Krishna, the Hindu deity who advises a major character in the Hindu epic the Bhagavad Gita. The idea is that people can ask this “AI-powered spiritual companion” for advice. But journalists quickly realized that, lacking a filter, these chatbots started spitting out casteist and misogynistic responses. The chatbot even said that it’s acceptable to kill if one’s duty demands it. Experts worry that users could take these messages seriously if they believe they’re coming from a divine figure, and that people could weaponize this pattern to drive harmful agendas.
🚏🌡 Ocean temperatures hit record highs, and El Niño is coming to warm them further
Sea-surface temperatures hit record highs in April, beating the previous record set in 2016. Scientists attribute this warmup to the end of La Niña, a climatic condition that generally cools the seas; La Niña had been going on for 3 years but ended in March. We’re thus experiencing a return to “normal” after a relatively cool period. Ominously, meteorologists think there’s an 80% chance of the ocean-warming El Niño starting between July and September of this year.
🚏🍪 Scientists invented an edible, non-toxic battery
A team from the Italian Institute of Technology has unveiled a battery made entirely of edible, non-toxic components, using such ingredients as riboflavin (Vitamin B2), beeswax, seaweed, and activated charcoal. While most batteries can cause water electrolysis if swallowed (which damages stomach tissue), these batteries use a low enough voltage to avoid this danger. Edible batteries could be highly useful for kids’ and pets’ toys (where swallowing is always a big risk), as well as medical devices like drug-delivery robots that swim around in the bloodstream.
📖⏳ Worth your time
Some especially insightful pieces we’ve read, watched, and listened to recently.
How Nintendo Solved Zelda’s Open World Problem (Game Maker’s Toolkit) — Explores how the designers of The Legend of Zelda: Breath of the Wild, a huge open-world game, managed to organically guide players toward key points of interest without making them feel like they were being forced to follow a linear path. Also discusses the developers’ “triangle rule” for creating a landscape that hides interesting landmarks behind every corner. The talk is especially timely given that BotW’s sequel comes out this week.
Prompt Injection Explained (Simon Willison) — Introduces an underappreciated risk vector for companies using Large Language Models: if you insert user input directly into an LLM prompt, malicious users can write prompts that hijack the AI's behavior, in a similar vein to SQL injection attacks. Willison then introduces potential solutions, including a “dual language model pattern” that quarantines untrusted input into an LLM that can’t access the rest of the system.
Electric Buses Are the Future. Agencies Are Still Right to Be Cautious. (Transit Center) — Argues that battery electric buses, while superior to diesel buses in many ways, aren’t simply drop-in replacements. Transit agencies often lack the infrastructure to charge buses, from power lines to charging stations; mechanics may not be trained on how to repair electric buses; and scheduling can become fiendishly difficult, since charging electric buses takes far longer than refueling diesel ones. “Electric is the future, and agencies should be making plans for it,” but the changeover can’t happen overnight.
Britain Is Dead (Palladium Magazine) — A harsh but incisive diagnosis of how the UK’s institutions, economy, and state power have been slowly fading for the past century-plus. A major structural reason was that the British economy industrialized early and thus didn’t have to double down on science and technology to stay competitive; the economy instead became “over-geared toward non-productive economic rent-seeking.” The author argues that this ultimately led to a British elite that lacks technical skills and a civil service that lacks competent staffers.
The Tomato Harvester (Boom California) — A well-written account of California's rapid tomato farming transition from small individual farms to large industrial farming in the 1960s. The tomato and the harvester co-evolved (only certain kinds of tomatoes could survive getting threshed, and only some machines could keep the fruits intact), and the domestication of both plant and machine was due to the human animal.
🔍🚘 Lens of the week
Introducing new ways to see the world and new tools to add to your mental arsenal.
This week’s lens: Crumple Zones.
What happens when we hit the edge of what a system was designed to do? Usually, we experience some sort of failure mode. However, the nature of that failure can vary. We might experience graceful degradation, where failure is gradual and the impact distributed, or we might experience catastrophic failure, where we hit a hard wall. What causes the difference between the two? One factor is how well the system is able to absorb the energy created by hitting up against a system boundary.
When we design systems, they implicitly incorporate constraints: the boundaries that define the space where they work efficiently. In the course of working within these systems, we’ll inevitably come up against these constraints. When we do, we need a release valve to absorb the excess energy created by hitting the boundaries.
Human-centric systems should acknowledge these tendencies. Enter crumple zones. Cars are designed to spread a shock and protect the passengers of a vehicle. But crumple zones only work if many constraints are met: your height/weight are within range, you don’t have your legs on the dashboard, etc. However, when we are within the constraints the crumple zone was designed for, they dramatically increase safety for the people inside the vehicle.
Teams and organizations can also implement crumple zones. Allowing 20% time or creating space for low-stakes collaboration and play can provide ways for people to scratch an itch and work on the things that frustrate them. Providing forums for giving feedback to leadership and where leadership can provide meaningful answers lets people feel like their concerns are heard. These tools can provide ways for people to expend the energy of their frustration in a way that causes less harm (and, in the best case scenario, drives some improvements).
We should keep in mind that these tools are not uniformly positive. They can become an extra unofficial expectation to do things that are Good For The Company. They can become an excuse to not officially handle problems because people assume the crumple zone will handle it. They can become an excuse for people not to spend time on their primary responsibility. However, these things are not what crumple zones are for. A crumple zone is, by definition, not reliable and is prone to produce unlikely side effects. Airbags, for all their advantages, leave you with a face full of bruises.
We can think about the crumple zones in our own organizations. When people hit up against boundaries, do they have ways to safely dissipate the energy of that impact? On the flipside, are we too reliant on crumple zones, depending on them to handle things that should be within the realm of a system’s normal operation? By balancing availability and over-reliance, we can use crumple zones for what they’re good for: providing a safety mechanism at the boundaries of normal operation.
🔮📬 Postcard from the future
A ‘what if’ piece of speculative fiction about a possible future that could result from the systemic forces changing our world.
// The world of Generative AI is exploding. Simple text prompts are starting to generate increasingly complex images and textual answers. And now, video games are being generated by AI. What might the future hold as people come to want more and more personalized content?
// May 2027. A chat window at www.my-ai-game.com.
Welcome to the Q Emporium! If you’re looking for a personalized game, you’ve come to the right place!
Worried about price? Well, as they say, if you aren’t paying for it, you are the product! I kid, I kid… mostly. Listen, we’re upfront about how QEmp works. The stuff that helps make a customized game you love is the very same juicy data that improves our models.
For a one-time purchase of $500, we give you a game generated by a prompt like “Red Dead Redemption, but set in a cyberpunk world with airships and walkable cities.” For only $50 a month thereafter, we’ll keep the game up and running along with our basic content generation systems.
If you want something more alive and more customized to you, we’ve got you covered (for a small fee). Using only the feeds you authorize, we will craft an ever-evolving game filled with NPCs that quote your favorite fiction. By giving us access to your emails, social media accounts, and more, we can mine them for information to personalize those NPCs even further. And those are just the appetizers.
We’re excited to announce that, after years of training and development, we are also able to personalize your games based on your gaming and media consumption habits and history. Would you kindly let us scrape your Steam, Epic, Xbox, Playstation, and Nintendo accounts for your game catalogs, playtime, and achievements? Throw in your Netflix, Kindle, Audible, and Spotify accounts, and now we’re cooking with gas. By analyzing how you actually consume media and what plot lines, quests, and types of content you enjoy interacting with, we can build a far more exciting and engaging game for you.
Oh, right… cooking with gas was before your time.
Listen, we know you’re busy juggling your five gig jobs, so we’ll keep it brief: our premium package results in an ever-evolving game world, custom-built for you. Your friends will play with you. And if they can’t play, we’ll simulate them for you, but just the right amount.
Why let the limits of your imagination stop you from escaping the drabness of your reality? Sign up today! The Q Emporium awaits.
© 2023 The FLUX Collective. All rights reserved. Questions? Contact flux-collective@googlegroups.com.