🌀🗞 The FLUX Review, Ep. 161
September 12th, 2024
Episode 161 — September 12th, 2024 — Available at read.fluxcollective.org/p/161
Contributors to this issue: Neel Mehta, Wesley Beary, Boris Smus, Melanie Kahl, Ade Oshineye, Ben Mathes, Erika Rice Scherpelz, Gordon Brander
Additional insights from: Justin Quimby, Dimitri Glazkov, Alex Komoroske, Robinson Eaton, Spencer Pitman, Julka Almquist, Scott Schaffter, Lisie Lillianfeld, Samuel Arbesman, Dart Lindsley, Jon Lebensold, Kamran Hakiman, Chris Butler, MK
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.
“Premature legibility makes ideas understandable, but at the price of any true novelty.”
— Michael Nielsen
🍊🧞 Unpacking “Founder Mode”
Founder Mode, the subject of a viral Paul Graham essay, is the tech world’s idea du jour. (The gist: startup founders are told to enter “manager mode” to scale their companies, giving their experienced hires plenty of latitude — but Graham argues founders have to be more hands-on, remaining in “founder mode.”) As the discussion around founder mode shows, it’s a vague enough concept to become an inkblot for our fears and desires. Yet if we dig deeper, we might find a version that is more robust.
The seductive take leans on charismatic authority and heroism, echoing the Great Man theory of history. It feeds fears that middle managers — “professional fakers” — are covering up their failures. It’s a handy excuse for micromanagement. None of this scales. As Graham notes, founders need to avoid founder syndrome, where they are unable to delegate. And as Steve Jobs pointed out, truly gifted people don’t want to work for micromanagers.
Yet something about founder mode resonates. In the post-ZIRP era, we can no longer say that CEOs don’t or shouldn’t steer. Leadership matters more than ever, and founders can offer a comparative advantage. The founder mindset can build a culture that cuts through the thermocline of truth. So instead of looking at founder mode through the lens of who is at the helm, we can look at it through the lens of what effective founders do to intervene strategically and create a culture that scales. We can move from founder mode to the founder mindset.
Dive deep: Founders dig for the substance behind the status report. They are not satisfied with superficial overviews. They proactively provide corrective feedback without trying to call all the shots. They go to the gemba, literally or through expectations such as all leaders regularly having skip-level 1:1s. Or, like Brian Chesky, their embrace of founder mode includes imposing a system of reviews.
Foster psychological safety: Although sometimes erroneously associated with low standards and coddling people, psychological safety is about building a culture where people can take risks. The test is to ask “did the last person who gave me bad news regret it?” When we don’t know (or don’t care) about this test, people tend to optimize for keeping us happy instead of achieving ambitious goals. As Michael Carey puts it, “if you kill the messenger, you just get less mail.”
Scale our own vision: Founders need to scale themselves, not become the bottleneck. Graham is right when he points to the harm of founders succumbing to generic leadership advice. However, the solution is not to hold the vision tightly, operating like the Eye of Sauron (the subject of this week’s Lens) and disempowering people. In that world, anything the founder is not currently watching fails. Scaling the founder’s vision requires figuring out how to replicate it virally. We have to train and empower people so that they eventually exceed our capabilities while remaining true to our vision.
A good leader acts as a floor, not a ceiling. They understand that they need to find people who are better than themselves in key areas and give them both autonomy to innovate and strong feedback to keep them aligned to the vision. This is why the best sports teams are run by people who know that they can’t be on the pitch at the critical juncture — but must find ways to decentralize command so that someone else can make good decisions about when to pass and when to shoot.
Our advice: escape the false dichotomy between founder mode and manager mode. Instead, accept that even Steve Jobs said, “we hire smart people so they can tell us what to do.” Your role as a leader is to define a direction with clarity and then build systems that get the best out of your people so you can get to a compelling destination together.
🛣️🚩 Signposts
Clues that point to where our changing world might lead us.
🚏🏜️ Roblox will let you generate 3D environments from a text prompt with AI
The video game platform Roblox plans to unveil a new generative AI model that can build full 3D scenes from just a text prompt (e.g. “generate a race track in the desert”); it’ll also be able to alter scenes given user feedback. Roblox did this by “tokenizing” each type of 3D block that makes up its virtual worlds and having the model place “tokens” based on what makes the most sense in context — analogous to how LLMs predict the next linguistic “token” in a sequence. A separate model will evaluate the generated 3D environment and check for oddities or inconsistencies, in which case it will force the 3D model to regenerate.
🚏🚕 Kenyan Uber drivers are ignoring Uber’s rates and setting their own
Uber drivers in Kenya have complained to Uber about low earnings to no avail, so an organization of 15,000 drivers has been circulating a new suggested fare schedule with prices about 50% higher than Uber’s posted ones. Passengers are seemingly still able to book via the Uber app but will be encouraged to pay via cash or Kenya’s M-Pesa mobile payment system. (One imagines that drivers won’t have to pay a cut of their earnings to Uber with this method, either.) While drivers have praised the collective action, some riders are frustrated at having to negotiate above the sticker price.
🚏🇷🇺 Russia wants to spend $646 million to block VPNs
Russia’s communication and censorship agency, Roskomnadzor, plans to spend $646 million to “upgrade its internet traffic-filtering capabilities,” which could reportedly be used to block or slow down VPN traffic. (Many foreign sites have been forced out of Russia since the Ukrainian invasion, so many Russians have turned to VPNs to access them.) The move comes five years after Russia passed a law to achieve what it called “digital sovereignty” — in other words, cutting itself off completely from the global internet.
🚏👨🚀 A billionaire will conduct the first private spacewalk
SpaceX has launched a billionaire tech entrepreneur 870 miles (1400 km) above the Earth’s surface — far above the International Space Station and farther than any astronaut (besides the moon-bound ones) has gone from our planet. He and a crew member will don spacesuits and briefly step out of the spacecraft. If it succeeds, it’ll mark the first time private citizens have conducted a spacewalk.
📖⏳ Worth your time
Some especially insightful pieces we’ve read, watched, and listened to recently.
The Slow Evaporation of the Free/Open Source Surplus (Baldur Bjarnason) — Argues that continued layoffs and reduced investment are reducing the amount of excess developer capacity that has historically gone toward open-source development. Combined with a hyper-focus on AI investment, which uses relatively little open source, this may mean that we’re in for a correction.
Preliminary Notes on the Delvish Dialect (Bruce Sterling) — A tongue-in-cheek piece about the distinctively stilted writing style of LLMs, which forms “the world’s first patois of nonhuman origin.” (Some calling cards: “delving into” things; strange analogies; “whether you’re X or Y”; “in conclusion”; “in the dynamic world of.”) This distinctiveness can obviously be used to identify and downrank AI-written text (as in the case of search engines), but it could also help intentionally find and promote AI text — consider someone looking for non-human text that’s guaranteed to be agreeable, anodyne, and reasonably helpful.
Your Customers Hate MVPs. Make a SLC Instead (Jason Cohen) — Argues that MVPs are too Minimal and rarely Viable. Instead, Cohen pleads, make sure that your first stab at a product is simple, lovable, and complete: “customers want to use a v1 of something simple, not v0.1 of something broken.” The floor for a SLC is higher than that of an MVP, too: “A SLC that never gets additional investment is a good, if modest product.”
I Tried Simulating The Entire Ocean (Acerola) — Showcases the incredible mathematical difficulty of simulating large bodies of water in real-time and describes how some of the most famous concepts in math (including the Fast Fourier Transform and Euler’s Formula) help make it work. The final rendering looks even more impressive once you realize how much work went into it.
🔍🪬 Lens of the week
Introducing new ways to see the world and new tools to add to your mental arsenal.
This week’s lens: the Eye of Sauron.
Imagine parents who are always watching. They believe their child will fail unless every aspect of their life is supervised. The parents constantly monitor, constantly offer advice and criticism. Their child grows up anxious, uncertain about their own judgment, and overly dependent on their parents — or anyone else.
The Eye of Sauron refers to a dynamic where an authority figure’s hyper-focus on a person or project leads to excessive control or oversight, leading to disempowerment and, often, fear. The leader often believes that they are helping a project that is on the path to failure, but their constant supervision fosters an environment of dependency, stifling growth and autonomy.
When leaders operate in this mode, they will find that they are shifting the burden of the problem to themselves. Those who are subject to the Eye become risk-averse and defer their judgment to the authority figure. People fear the Eye but become dependent on it: the leader will become a bottleneck as people constantly turn to them for approval.
What’s worse is that the Eye of Sauron doesn’t scale and can’t be everywhere at once. This creates situations where parts of the organization that are currently unmonitored are unable to perform or improve until the Eye’s gaze returns to them. As the organization or family grows larger, these neglected areas develop problems, which the Eye uses as justification for further micromanagement. This thus creates a reinforcing feedback loop: “I have to watch everything and micromanage everything because things don’t happen if I don’t do it.”
The Eye of Sauron is sticky: once experienced for a little while, it becomes an organizational pathology, becoming more and more difficult to disengage from over time. The risk aversion creates a sense of apathy when viewed from above, and the most logical solution appears to exert more oversight, creating a vicious cycle.
Rather than focusing on constant oversight, leaders should focus on building trust, setting clear expectations, and giving frequent feedback. The key is to redirect from control to empowerment, but do so in a way that maintains alignment and high standards.
© 2024 The FLUX Collective. All rights reserved. Questions? Contact flux-collective@googlegroups.com.