Simple Systems Lose Wars

Or, Why ‘Keep It Simple’ Is Stupid

Everybody loves simplicity.

Engineers lust after it. After all, complex systems multiply points of vulnerability or failure. Simple systems are easier to maintain and harder to break.

In business, clarity wins. If your executives, employees, and customers can’t summarize your value in a sentence, you have a problem. That’s why marketers craft “elevator pitches.”

In war, simplicity can lose you everything. War is not like business.

In business, firms compete side by side, like runners in a race. They do not shoot at each other. As strategist Edward Luttwak has observed, in war every move provokes a countermove from an intelligent adversary. Markets may ignore you. Enemies do not.

In war, simplicity becomes predictability; predictability invites penetration and points to countermeasures.

Consider Israel’s air defenses. Its adversaries chose a seemingly rational strategy: saturate the air with swarms of relatively simple rockets and drones to overwhelm the defense.

That might have worked well against a single defensive layer. Facing swarms of rockets taking low altitude paths, Iron Dome maybe could be saturated. But the enemy faced Iron Dome + David’s Sling + Arrow + Iron Beam + advanced distributed sensors + AI triage. To succeed, an attacker would have to vary altitude, speed, and vectors, and overwhelm multiple layers.

Hardest of all would be to defeat AI triage, the algorithmic allocation of defense measures. This would require not merely response but anticipation of multiple possible responses. With AI, the defender is no longer simply absorbing blows. It is adapting, raising the attacker’s cost nonlinearly.

In our time, the advantage of complexity is spiraling upward because AI manages complexity beyond human cognitive limits. It integrates distributed sensors, allocates resources, predicts trajectories, and adjusts responses in milliseconds. AI transforms complexity—multiple options—into adaptability. The more adaptive a defense system becomes, the harder it is to overwhelm economically.

Here is the paradox. Internally, good systems are elegant and, yes, even simple. Externally, good defenses are confounding—because of complexity.

The internet offers a civilian analogy. Internally, routers follow relatively simple rules. But the path taken by a single message can be astonishingly complex compared to the dedicated circuit of the old telephone system. Those complex routes allow the network to route around damage; it allows for adaptive resilience.

For investors, the lesson extends far beyond missile defense.

In markets today, the winners increasingly offer not products but systems that integrate hardware, software, networks, feedback, and energy sources. None of the “Magnificent Seven” are product companies. They are system companies.

Nvidia (NASDAQ: NVDA) may appear to compete with Intel (NASDAQ: INTC) and AMD (NASDAQ: AMD). But CUDA—NVDA’s software development platform—makes that comparison almost quaint. Nvidia orchestrates an ecosystem: developers, software libraries, networking fabrics, and tightly integrated hardware. The adaptive advantage lies not in the silicon alone, but in the architecture that allows programmers to extend, optimize, and refine the system continuously.

We used to analogize a business’s enduring advantages to a moat or a wall. Heavy and static, anchored by “replacement cost.” Today, enduring advantage comes from complex adaptive systems that would be hard to describe, let alone replace.

Today, the great simplifiers in our world—the expert simpletons—are focusing on energy as the source of supremacy in AI. Their simple formula: the side that wins in energy will fuel the largest datacenters and will have the most tokens and parameters and thus will have the best AI. The best AI will achieve the most advanced industry and military applications.

But as paramount physicist Richard Feynman, a man we quote regularly in Gilder’s Technology Report, put it: “Science is the belief in the ignorance of experts.”

More likely, the country that wins in energy will fall for the illusion that intelligence consists in tokens and parameters of old expertise, enshrined in multi-gigawatt data centers rather than in 12-to-14-watt human brains.

The winners will be the countries and companies that endow and emancipate human minds rather than those who mistake simple power generation in machines as the secret of the complex surprises of real—human—generative intelligence.

At Gilder’s Technology Report, we look for technologies that exploit science to alter cost curves and create new markets. When raw power and mass are replaced by intelligence, entire sectors can be revalued—rapidly. Don’t blink.