🌀🗞 The FLUX Review, Ep. 145
May 9th, 2024
Episode 145 — May 9th, 2024 — Available at read.fluxcollective.org/p/145
Contributors to this issue: Neel Mehta, Boris Smus, Ade Oshineye, Ben Mathes, Erika Rice Scherpelz, Dimitri Glazkov, Jon Lebensold, MK
Additional insights from: Justin Quimby, Alex Komoroske, Robinson Eaton, Spencer Pitman, Julka Almquist, Scott Schaffter, Lisie Lillianfeld, Samuel Arbesman, Dart Lindsley, 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.
“When the facts change, I change my mind — what do you do, sir?”
― John Maynard Keynes (attributed), in response to criticisms that he changed his opinions frequently
🐜🐜🐜 The trends go marching one-by-one, hurrah, hurrah
In biology, an ant mill is a phenomenon where ants lose their path and instead just follow each other around the same loop until they die. Ant mills are a good metaphor for business trends, where companies start doing something, perhaps something well-founded to start, which becomes a trend with its own life. Many of the layoffs in Silicon Valley in the past couple of years have the feel of an ant mill.
Ant mills gain their power from feedback loops. There is a reinforcing feedback loop: as ants follow each other's trails, they lay down pheromones to reinforce that trail. If the trail is in a circle, that path can become strong quickly. A balancing feedback loop makes it worse: anything outside of the trodden path becomes weaker and weaker over time until it dies away.
Beyond macro trends, we can also see ant mill dynamics underlying Conway’s Law, the idea that organizations tend to produce products that mirror their internal structure. When one first encounters Conway’s Law, it seems like it ought to be easy to escape, but after a while, it seems inevitable. That’s because, like with the ants, there are two intersecting feedback loops that connect organizational charts with the products a company is able to produce:
A reinforcing loop that rewards you for structuring your product in alignment with the org chart.
A balancing loop that punishes you for structuring your product in ways that don’t match the org chart.
For example, if you are building software that matches the structure of your organization, you will have an easier time getting support and staffing. If your goal is misaligned, you will have to meet with N different leaders. Even if you get something started, your project is at high odds of cancellation because it’s not quite aligned with the goals of any one of those leaders.
We often write about the sophistication and merits of collective intelligence and self-organization. The same group dynamics that lead to ant mills can lead to complex but positive phenomena such as ant rafts. However, we must also keep in mind that blindly responding to stimuli doesn’t always lead to good outcomes—either for the individual or the collective. We have to fight to retain the agency that exists in the gap between stimulus and response so that we aren’t completely at the mercy of incentives.
Without that agency, it’s hard to escape the ant mill. This applies whether it’s the organizational ant mill of Conway’s Law, the cultural ant mill of being caught up in the issues of the day, or any other situation where you are rewarded for going along with the flow and punished for deviating. Escaping requires the ability and willingness to resist the incentives that surround you. Or, perhaps, if you are in a position to shape the incentives — and trust yourself to know what a good outcome actually looks like — you can pull an inverse Conway maneuver and shape the structural incentives to change the resulting behavior.
🛣️🚩 Signposts
Clues that point to where our changing world might lead us.
🚏🇺🇦 Ukraine’s foreign ministry is launching an AI spokesperson
Ukraine recently unveiled an AI-generated avatar that would “make official statements on behalf of its foreign ministry.” While the statements will still be written by humans, the digital character will read the statements on video and move ‘her’ hands accordingly. To prove the authenticity of these videos, Ukraine will pair the videos with QR codes that link to textual versions of the statements on the foreign ministry’s website.
🚏🧯 Colorado is buying firefighting helicopters so it doesn’t have to share with other states
States in the western US have traditionally shared firefighting helicopters with each other to help combat the increasingly frequent (and increasingly devastating) wildfires in the region. But Colorado’s governor and fire chief have announced a $6.9 million plan to buy two helicopters, fuel trucks, and other equipment, saying that this mutual aid model would no longer work: “[when] there are fires in California and fires in New Mexico… resources are deployed there, and there aren’t any left for Colorado when we need them.”
🚏🚇 A new hack could “neuter the entire purpose” of VPNs
Researchers found a vulnerability called TunnelVision that would let the owners of internet networks (like public Wi-Fi networks at coffee shops or airports) trick VPNs into sending all traffic over an unsecured tunnel that the attacker could snoop on, completely bypassing the “secure tunnel” that VPNs offer to maintain privacy and security. The only operating system that’s immune to the hack is Android, which doesn’t implement the particular networking feature (DHCP Option 121) that the hack relies on. (You could also run the VPN on a virtual machine that’s not directly connected to the computer’s networking stack.)
🚏🛜 An ISP got a $206 million loan backed by IP addresses
Cogent Communications, an American internet service provider, announced that it’d be selling $206 million in secured notes (a sort of loan backed by collateral). This is normal enough, but the unusual bit was that their collateral was a block of IPv4 addresses; sources estimate that Cogent has about 10 million addresses. There are only 4.3 billion possible IPv4 addresses, so they’re getting both scarce and expensive — Amazon Web Services started charging for them last year, noting that their price had tripled in the past five years — and Cogent’s loan goes to show just how valuable they’ve become as assets. (There’s a virtually unlimited pool of the newer IPv6 addresses, but they aren’t as widely supported yet.)
📖⏳ Worth your time
Some especially insightful pieces we’ve read, watched, and listened to recently.
What Can LLMs Never Do? (Strange Loop Canon) — Observes that LLMs struggle with evaluating formal systems like Conway’s Game of Life or solving multi-step puzzles like Wordle. From this, the author argues that LLMs are great at applying patterns they’ve seen in their dataset but poor at generalizing to find novel patterns or applying logical operations in sequence. In short, these machines “demonstrate incredible intuition but limited intelligence.”
The Corrosive Downside of Acquihires (Mark Suster) — Explores the downsides of acquihires for the acquiring company. The acquired founders are likely to leave quickly for their new venture, while existing employees resent the newly joined cohort for their higher salaries and contemplate founding their own startup.
Of Course America Fell for Liquid Death (The Atlantic) — Evaluates the surprising success of Liquid Death, a brand of pricey bottled water whose primary differentiation is its edgy and stylish bottling. While for many companies the brand has become more important than the product itself, Liquid Death represents the “apotheosis” of this trend: there’s basically no product at all, and the brand and marketing drive themselves. In a sense, the advertising is the product people are lining up to buy.
The Year That Killed Received Pronunciation (Geoff Lindsey) — Describes how the social and cultural shifts of the ‘60s turned Britain’s most prestigious accent (nicknamed the “Queen’s English”) into a target of derision and mockery, as working-class and regional accents started to gain popularity. This surprising reversal mirrored broader cultural trends of the era.
🔍📆 Lens of the week
Introducing new ways to see the world and new tools to add to your mental arsenal.
This week’s lens: tourists and residents.
Imagine yourself in Indonesia, Senegal, or France. Do you imagine yourself as a tourist—exploring, discovering, maybe a little uncomfortable? Or do you imagine yourself as a resident—comfortable, efficient, maybe a little complacent? As residents, we are familiar with the norms of the culture and the lay of the land. As tourists, we bring a different perspective and are actively looking to have a new experience.
When we springboard into a new role, a new group, a new hobby, or a new organization, we start off as tourists. Whatever norms and questions we were comfortable with would change to some degree. This isn’t a bad thing. The reason we engage in change often is motivated by the desire to expand our horizons. Many folks do post-undergraduate degrees in a different field with the understanding and goal of shifting from their established environment to becoming a tourist and learning something new. That something new may become a career change and new life journey, or they may take their new insights back “home” and reap the benefits of cross-pollination.
It’s valuable to spend time in our lives as both tourists and residents. It’s also valuable to think about the distribution of residents and tourists within a particular group. Residents know an area with a depth that a tourist cannot replicate. They are the deep domain experts who can solve problems that no one else can. But if residents never get an infusion of new ideas and insights from tourists, they can become complacent. Ideally, there is a mix of residents and tourists, with any particular person taking on different roles at different times.
We can apply this ‘tourists and residents’ lens in our own lives. At a personal level, we can look at what part of our lives we reside in as residents and which we relate to as tourists and see if we are hitting the right balance. For organizations we are a part of, we can look at if we have the right balance. Then, if we find discrepancies, we can find ways to invest in balance, knowing that both modalities bring value.
© 2024 The FLUX Collective. All rights reserved. Questions? Contact flux-collective@googlegroups.com.
The ant mill / Conway's story reminds me of the joke about meetings, teamwork, collaboration: "none of us is as dumb as all of us."
Of course collective intelligence is capable of amazing things, too. I think what this really points to is that the individual and the collective each have a set of incentives that aren't always aligned. Your behavior in the group won't necessarily match your tendencies as an individual.