🌀🗞 The FLUX Review, Ep. 181

March 13th, 2025

Mar 14, 2025

“FCP-181” // Photo: MK with Midjourney and Claude 3.7 Sonnet

Episode 181 — March 13th, 2025 — Available at read.fluxcollective.org/p/181

Contributors to this issue: Neel Mehta, Julka Almquist, Erika Rice Scherpelz, Steffano Mazzocchi, Spencer Pitman, Justin Quimby, MK

Additional insights from: Ade Oshineye, Alex Komoroske, Ben Mathes, Chris Butler, Dart Lindsley, Dimitri Glazkov, Jasen Robillard, Jon Lebensold, Kamran Hakiman, Lisie Lillianfeld, Melanie Kahl, Robinson Eaton, Samuel Arbesman, Scott Schaffter, Wesley Beary, Boris Smus

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.

“On two occasions I have been asked, ‘Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?’ I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.”
— Charles Babbage

➡️🎨 Shifting the craft

One useful way of thinking about AI so far has been AI as a floor-raiser. AI acts as an equalizer, giving more people access to knowledge and capabilities once reserved for experts. If you can describe a problem well enough, AI can generate a competent solution—at least in domains like writing, research, and coding. A tool that raises the floor allows more people to participate. It makes tasks easier or more intuitive and helps navigate unfamiliar domains (just be wary of Gell-Mann amnesia).

But there’s another shift underway, one that felt far out of reach even a year ago: AI as a ceiling-raiser, expanding what’s possible at the highest levels of skill. AI as a tool that raises the ceiling pushes the limits of what experts can do, amplifying skill rather than replacing it. We’re starting to see hints of this in AI-assisted coding, where tools like Claude Code using models like Claude 3.7 Sonnet are opening powerful new possibilities even for experienced programmers. These ceiling-raising tools still require skill to be wielded effectively, but the payoff is that they can make even experts more effective.

In this world, the nature of expertise changes. To stick with the example of code, writing great code will no longer be about knowing the ins and out of libraries and languages. But neither will writing great code with AI be just about prompting well. It’s about understanding how to break down problems, define success criteria, and critically evaluate the results. Planning and evaluation will be the new coding. The tool doesn’t replace the craft—it shifts it. Especially with the constant change in model capabilities, the opportunity isn’t in mastering any single tool, but in learning how to work with these tools in a deeper, more intentional way.

AI is not novel in its dual nature as a floor and ceiling raiser. Leadership, for example, can be treated as floor-raising: ensuring the team is equipped, the barriers are lowered, and the system works. But great leaders push the ceiling higher, challenging their teams to think differently and to engage with new approaches in ways that expand what’s possible.

Even equipped with an army of agents, we can’t just turn on the computer and tell, “Build me a video game!” We have to think through what we want out of it. It’s not about issuing commands but about conversing with the tool, shaping its output, refining the vision, and navigating tradeoffs. AI helps raise the floor for many, but for those willing to engage at a higher level, it’s also an invitation to raise the ceiling.

🛣️🚩 Signposts

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

🚏🕰️ American workers are clocking out 42 minutes earlier than before

New data shows that the average workday in the US now ends at 4:39pm, a full 42 minutes earlier than two years ago, when it ended at 5:21pm. Despite the shorter day, the report shows that productivity has increased by 2% in this same stretch of time. The number of people working on the weekends has increased slightly, though, which suggests that workers are more comfortable spreading their work throughout the week.

🚏🧳 Southwest Airlines is ending its ‘two free checked bags’ policy

“Bags fly free” was long one of Southwest Airlines’ signature policies, but the company recently announced that it would be ending the perk for all but a handful of high-status travelers. The stock price briefly spiked on the news, but analysts panned the move, saying “this is how you destroy a brand” and “we are watching an airline self-destruct… this is the equivalent of deliberately sailing a ship into an iceberg.” (Indeed, one analyst said that 97% of Southwest customers chose the airline for the free bags.) The end of this beloved policy is one of many moves the airline has made in response to “intense pressure from an activist investor”: introducing basic economy, ending open seating (and charging for seats with more legroom), launching red-eye flights, and implementing its first-ever layoffs.

🚏🇸🇾 Syria hosted its first international tech conference in 50 years

Last month, a group of Syrian-American technologists organized a conference in Damascus aiming to “connect Silicon Valley with Syria’s emerging tech ecosystem,” with a focus on finding ways to overcome skill and infrastructure gaps in the fledgling Syrian tech industry (which has spawned about two tech companies a year over the last decade). This was the first independent tech conference in Syria since the end of the civil war and, indeed, the first since the Assad regime took power half a century ago. The organizing group hopes to create 25,000 new jobs in Syria’s tech sector over the next five years.

🚏⌚ Smartwatch sales fell for the first time ever

Worldwide smartwatch sales fell by 7% in 2024 compared to the previous year, marking the first such decrease in smartwatch history. Analysts believe two major drivers were a burst in India’s smartwatch bubble as customers grew dissatisfied with a flood of cheap, low-quality devices, and North American customers did not see many new features from Apple watches. One expert observed that “the smartwatch has gone from being a new and exciting gadget, to something now that’s stabilising” — which explains the boom in smartwatch sales in China, where several companies compete in a price war (to the benefit of shoppers).

📖⏳ Worth your time

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

The Kaiser and a "Mediocre Man" Theory of History (Kiran Pfitzner) — Argues that both the “great man” theory of history and its antithesis (that history is primarily driven by structural factors, with limited agency available to individual actors) are flawed. The author stakes out a synthesis position: individual people thrust into positions of power make their mark on history, but many of them aren’t exceptionally competent. History is thus contingent on their weird decisions or failures. It’s, therefore, worth historians’ time to study history's many underexplored ‘mediocre men.’
The Tiny Star Explosions Powering Moore’s Law (IEEE Spectrum) — A microchip manufacturing scientist explains how the physics of supernova explosions (which he got interested in as a youth thanks to his grandfather, an amateur astronomer) helped his team figure out how to etch smaller-than-ever circuits on silicon chips. The science involved is awesome: laser beams blast a droplet of molten tin, which reaches 200,000 ºC and emits a burst of plasma that gets focused into a beam of light that draws on the silicon wafer. The same equations that predict massive supernova explosions also predict how the tiny clouds of tin plasma expand!
The Successful Presidents Curse (Brian Beutler) — Argues that many politicians try to imitate the policy positions of successful politicians past, when really those pols’ success was due to their unique charisma. Campaign consultants can’t change candidates’ raw talent (or lack thereof); policy is the only variable they can control—so it’s not too surprising that advisors over-focus on policy and don’t talk much about messaging style.
Why π^π^π^π Could Be an Integer (For All We Know) (Stand-Up Maths / YouTube) — The popular mathematician explores how the limits of computation and the dearth of mathematical knowledge of transcendental numbers make it impossible to know whether pi raised to itself three times is a rational number or not. It would seem impossible for a transcendental number to turn rational with just a few operations, but observe that, for irrational numbers, (√2)^(√2)^(√2) = 2.

🔍🥧 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 fixed pie.

Oh no! It’s another big corporate bankruptcy that splashes across headlines as lawyers, creditors, and investors scramble to salvage what’s left. After a bankruptcy, the pie is fixed. The game is about dividing the remains; every extra slice one stakeholder takes means someone else gets less. Contrast that with technological innovation, where success often compounds—a rising tide, a blue ocean, a bigger pie for all.

The fixed pie mindset is a well-known trap in negotiation theory: an anti-pattern where people assume resources are limited and the only way to win is to take from others. It leads to zero-sum thinking and missed opportunities for value creation. When you see the world as a fixed pie, whatever slice you can grab is all there is—and the most generous, win-win scenario is a 50/50 slice down the middle. By contrast, value creation strategies aim to grow the pie through innovation, collaboration, or opening new markets.

Problems arise when we apply fixed-pie thinking in places built for growth. In technological innovation, a scarcity mindset can turn potential partners into competitors and shrink possibilities before they even emerge. It also shows up as over-optimization—increasing your profits by cutting away what seems inessential, only to discover that what you trimmed was load-bearing. Scarcity thinking can create brittle systems that are over-tuned to prevent loss instead of designed for resilience.

On the flip side, acting like you’re in a blue ocean when you’re carving up the remains of a finite system—like a distressed asset or a saturated market—can lead to magical thinking and painful missteps. The skill is recognizing whether you're in a world of fixed constraints or one ripe for expansion.

Check your pie. Before you negotiate, collaborate, or strategize, pause to ask: Are we dividing what’s left, or can we create something new? The answer shapes not just how you play, but how much pie there is to go around.

🌀🗞 The FLUX Review