🌀🗞 The FLUX Review, Ep. 153

July 11th, 2024

Jul 12, 2024

Episode 153 — July 11th, 2024 — Available at read.fluxcollective.org/p/153

Contributors to this issue: Neel Mehta, Boris Smus, Erika Rice Scherpelz, Ade Oshineye, Spencer Pitman, Chris Butler, Ben Mathes, Justin Quimby, Melanie Kahl, MK

Additional insights from: Dimitri Glazkov, Alex Komoroske, Robinson Eaton, Julka Almquist, Scott Schaffter, Lisie Lillianfeld, Samuel Arbesman, Dart Lindsley, Jon Lebensold, Kamran Hakiman

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.

“I used to argue the matter at first, but I’m wiser now. Facts are stubborn things, but as some one has wisely said, not half so stubborn as fallacies.”
— L.M. Montgomery

👁️📈 Leading beyond the data

Data-driven decision-making is crucial. Data itself cannot make a decision, creating a paradox for leaders. If we blindly follow the data, we abdicate our responsibility to lead. However, we can’t follow only our intuition—not if we want to make good decisions.

But is this really a paradox? “Don’t use data, but also use data”?

No. Use data within a context and inform that context. Data must be filtered through a model such as our values, experiences, tastes, or worldviews. Then, to prevent overfitting, our interpretation of data and the models we filter them through must be balanced with empiricism, hypothesis generation, experimental design, and feedback loops.

Even if we set up simple rules — “If sales are above $X this month, invest more in marketing”, or “If my partner comes home after 10pm five times this week, ask them where they have been” — the data isn’t deciding. We set up the conditional. The data informs and influences us, but we choose based on the totality of our experience. We generally start by defining these rules from our empirical knowledge. This ensures that our models are grounded in reality.

However, experience alone does not allow us to predict. For that, we need to use the tools of hypothesis generation and experimental design: What do we think the levers in the system are? How would we test that? How would we detect if we are wrong? The data acts within our context to validate or discredit our hypotheses.

Finally, we must be aware of balancing and reinforcing feedback loops. When we are more than mere observers, when our actions impact the world around us, we risk our models outrunning the reality that inspired them. And this is a risk. Systems without effective balancing feedback loops eventually run amok. This requires us to pay attention to leading and lagging indicators in the data to help us decide whether the system’s behavior needs to be balanced or reinforced.

The financial sector provides rich examples of models run amok and ways to balance feedback loops. Leveraged buyouts and mortgage-backed securities were valuable financial tools, but they became problematic when the reinforcing feedback loops of their overuse caused broader repercussions in the financial system. The much-discussed idea of bonus clawbacks solves the problem of rewarding people based on (mis)leading indicators.

Rewards and incentives bring us to gamification. Gamification brings game mechanics such as competition and scorekeeping to activities that are not traditionally games. Gamification can shape behavior and provide visibility… but it does so by incentivizing people to play the game rather than solve the problem. It’s a valuable tool in the toolbox, but we must consider that the game can eclipse reality, as happened with the 2021 GameStop short-selling.

To navigate the complexity of the real world, we don’t blindly follow data. Instead, it’s a dance of expertise, data, modeling, feedback, and ever-more-plausible explanations. Especially when our actions can shape the very world our models aim to describe, avoiding getting stuck in our existing model creates opportunities and resilience.

🛣️🚩 Signposts

Clues that point to where our changing world might lead us.

🚏⛈️ Cloudflare says it’ll let websites block AI scrapers that ignore robots.txt

AI web scrapers are supposed to honor websites’ robots.txt requests not to scrape them, but some AI companies have been accused of ignoring the honor system and scraping anyway. So, Cloudflare is announcing a tool that it claims can identify these rogue AI scrapers and block them from accessing the websites that Cloudflare protects.

🚏🇬🇧 The UK became the world’s #2 economy for startup funding

Startups in the United Kingdom raised a collective $6.7 billion in the first half of 2024, overtaking China (which saw $6.1 billion in investment) to reach the global #2 spot for startup funding. (The USA’s $54.8 billion in funding during the same timeframe made it the #1 startup economy by far.) However, experts warn that Britain has also suffered from a lack of unicorns, a drought of IPOs, and a decline in foreign investment.

🚏🤑 US Senators’ new proposal would block Congress members from trading stocks

A bipartisan group of American Senators introduced a bill that would prevent members of Congress (plus their spouses and dependent children), the President, and the VP from buying or selling individual stocks while in office. (It’s unclear what forms of investment would still be allowed.) The phenomenon of Congress members making profitable trades due to the material nonpublic information they get has become a running joke over the years, and financial firms have launched ETFs that follow the politicians’ stock trades.

🚏🇳🇱 Half of Dutch gas stations are expected to close in the next decade

According to projections from one research firm, half of the Netherlands’ gas stations will likely close “in the next five to ten years,” going from 4,131 stations today down to about 2000. The decline in petrol will be mainly due to the rise of electric cars. Small, independently-run shops are most likely to close, with the research firm expecting a “spate” of consolidation and takeovers as the industry contracts.

📖⏳ Worth your time

Some especially insightful pieces we’ve read, watched, and listened to recently.

We’ll Never Have a Model of an AI Major-General: Artificial Intelligence, Command Decisions, and Kitsch Visions of War (Journal of Strategic Studies) — Argues that commanders in war need to use abductive logic (judgment and decision-making based on context) while AIs are only capable of inductive logic (probabilistic inference based only on known facts). While machines can play ‘war’ games like Chess or Go well, it’s folly to compare that to real war: you never have perfect information, closed systems, and frictionless feedback loops in real combat.
Do Chairs Exist? (Vsauce) — In Michael’s trademark madcap style, he explores some of ontology’s most mind-bending questions: are descriptors like “chair” something above and beyond the components that make them up? If so, where does that distinction come from? Where are the boundaries between objects? What exactly defines a category? This deconstruction is fascinating from a systems-thinking point of view, given how much we talk about emergent behaviors and the whole being more than the sum of its parts.
Yimboree (Interfluidity) — Stakes out a nuanced position on how to build more housing. While the author agrees with YIMBYs’ goals, he disagrees with their methods: some policies (like land use reform) will only drive marginal improvements, while others (like getting homeowners to agree to densify their neighborhoods) may be unnecessarily confrontational and lead to pushback. The author’s preferred solution is to embrace greenfield development, but ensure that it’s dense.
Council of the Pecans (Orion Magazine) — Botanist and author Robin Wall Kimmerer writes vividly about the mysterious phenomenon of mast years, in which all trees in an area spontaneously produce a large number of seeds: “If one tree fruits, they all fruit — there are no soloists”, imploring us to learn wisdom from our natural ecosystem.

🔍🫂 Lens of the week

Introducing new ways to see the world and new tools to add to your mental arsenal.

This week’s lens: care-ing capacity.

In ecology, carrying capacity is the number of living things a habitat can support without environmental degradation. When an environment exceeds its carrying capacity, it degrades, and populations can suddenly collapse, seemingly without a specific cause.

When we’re constantly exposed to the world's troubles — personal, global, and at every scale in between — we find so much to care about. This can be enlarging, creating a sense of connection. However, this constant barrage can also become too much.

While being emotionally overwhelmed has always been part of the human experience, the interconnectedness and information influx means that, as a society, we may be hitting a moment of care-ing capacity. We’re hitting the edges of what we can care about and for before hitting personal degradation. For some, degradation may result in an obsession with the ills that overwhelm them. For others, it can lead to numbness or denial of those ills.

Extending this metaphor, this personal “care-ing capacity” intersects with the very real care crisis. The extension of care, both emotional and practical, to our intimate realm of children, sick people, neighbors in need, and elders, stretches our baseline capacity. We’re at care overload — and something’s got to give.

In the face of reaching our care-ing capacity, we need to be aware of the ways we can expand or modulate our capacity to care without falling into patterns of degradation. Epistemic hygiene tripwires can be valuable: look out for headlines or campaigns that intentionally aim to trigger our emotions. We can also look to modulate our concern through our agency: we can allocate more of our caring budget to things we can do something about, even if it is something small. Additionally, we can seek ways to care together and care proximately — strengthening our “care-y-ing” muscles through emotional or practical solidarity.

It can feel odd to hear that the way to maintain our capacity to care is to care less. However, like with any renewable resource, the way to make it last indefinitely is to be aware of the rate at which it is used and renewed. Knowing our personal care-ing capacity is what will allow us to maximize care in the long term.

(...Or at least until our systems carry — and care-y — more of the burden.)

🌀🗞 The FLUX Review