You’ve been appointed as Minister of Technology. Everyone in the conference room holds perfect strategy documents, dense execution paths, clear KPIs. You feel slightly uneasy but begin approving: “Let’s do this.”

Six months later, you find all plans are on track, but nothing exciting is happening. Damn it, you can’t explain why—all the metrics are green.

Kenneth Stanley and Joel Lehman use a programming example to illustrate why. They ask: If you want to evolve an AI that can pass through a completely specific goal (like navigating a maze fastest), what’s the most effective method?

The intuitive answer: directly optimize for this goal. Select those mutations that move closer to the finish line with each step.

In reality? These directly optimizing AIs often get stuck in local optima, never finding the real exit. Because each step moves in a seemingly correct direction, there’s no motivation to explore those “seemingly circuitous” areas—yet it’s precisely those circuitous areas that can bypass dead ends.

Now reverse it. What if you don’t directly optimize “getting out of the maze,” but instead optimize “exploring the most new states” (Novelty Search)?

Something strange happens. The AI starts wandering randomly. It enters dead ends, then exits. It explores corners, corridors, all sorts of lengthy routes. In pursuing “novelty,” it accidentally discovers a direct path to the exit—a path it would never have found if directly optimizing.

This experiment later changed Stanley’s view of human history.

The Turning Points in the Story

“Why Greatness Cannot Be Planned” uses historical cases to validate Novelty Search. The case selection is unbeatable:

The Wright Brothers. They wanted to build airplanes, but spent time not on engine engineering, but in wind tunnel experiments studying wing shapes and airflow behavior. Those seemingly completely unrelated experiments became the foundation for their later ability to control flight. Each “deviation from plan” experiment was a stepping stone to later success.

The Microwave. Percy Spencer, working at a radar company, discovered that microwaves could melt chocolate. That wasn’t his goal. The goal was developing better radar. But this “accidental discovery” later changed every household kitchen.

GPU. The video game industry needed faster graphics processing, driving GPU development. No one in the 1990s said: “We need GPUs to train large neural networks.” But those chips optimized for gaming later became the infrastructure for deep learning.

Common thread? All achievers were pursuing some point of interest, then were led in completely different directions. Each “deviation” wasn’t waste, but was constructing a space of possibilities.

This isn’t feel-good rhetoric. This is a mathematically verifiable pattern: Systems using Novelty Search often discover more powerful solutions than systems using Goal Search.

From Policy to Personal

The book’s middle section contains a disturbing chapter: education policy.

Governments worldwide are doing the same thing—defining “success,” then designing education systems to maximize this success metric. Test scores, college admission rates, employment rates. Each is measurable, so it seems scientific.

But what’s the result? They’ve cultivated a generation skilled at walking pre-planned routes, but lacking the ability to innovate when encountering problems without preset paths.

Stanley’s argument: If you really want to educate creative people, you need to let them explore. Let them “waste time” on seemingly unrelated projects. Let them pursue quirky interests, even if these interests aren’t in the curriculum.

This completely contradicts modern corporate doctrine. Modern enterprises say: focus. Set goals. Every minute should advance toward the goal.

Stanley’s response: That kind of focus traps you in local optima, never to escape.

My Own Three Stepping Stones

Reading this book, I was re-examining my own twenty-year entrepreneurial trajectory. Honestly, I didn’t start with a “five-year plan.” I had some intuitions and interests.

First Stepping Stone: iShelly. Online financial media. Strictly speaking, this was unlike anything I later did. But what did I learn in that process? Web traffic, content marketing, community management. These skills were used in every subsequent project.

Second Stepping Stone: nvesto. From financial media to investment consulting. On the surface it seemed like “deepening” the financial field, but actually I was exploring a completely different business model—from ad-driven to consultant-driven. It failed, but that failure taught me the real cost of client trust, and why professional services can’t operate with media logic.

Third Stepping Stone: Good Food Machine, Thickness Market, Half-Acre Pond. Suddenly jumping to agricultural food ecosystems. It seemed like wasting the expertise I’d accumulated in finance. But in that process I understood what real supply chains are, what farmers’ real situations are, what it means to “change an ecosystem’s difficulty.”

Now: Our AI Platform. Circular economy digitization. Why can I see things others can’t? Because I’ve walked those circuitous paths. Finance×Agriculture×Architecture×AI. Each “digression” is the foundation for what I can see at this moment.

If I had “planned” twenty years ago to become a circular economy AI expert, I would have missed all these stepping stones. I would have been like a fake expert, knowing only surface knowledge, lacking real tactile understanding.

Why Organizations Become Death Mechanisms

Stanley’s sharpest criticism in the book: Goal-oriented systems naturally oppose novelty.

In highly goal-oriented organizations, people who “deviate from goals” are considered unprofessional. “Wasting time” on seemingly unrelated projects is seen as laziness. Cross-boundary experiments are viewed as distraction.

Yet it’s precisely in these activities despised by organizations that real innovation emerges.

So the more successful large organizations become, the more they set clear goals, the harder it becomes to produce game-changing innovation. They become highly optimized systems, perfectly solving past problems but with no defense against future problems.

This explains why Kodak engineers invented the world’s first digital camera as early as 1975, yet the company refused commercialization—not a technical problem, but a clear, definite goal (maximizing the film market) eliminated the motivation to explore new possibilities. They personally created the technology that would disrupt themselves, yet were blind to it due to goal-oriented thinking.

Treasure Hunter Mentality

Stanley uses a metaphor: goal-oriented people are measuring tapes. They know where the destination is, how far the straight-line distance is, they just want to arrive fastest.

Innovators are treasure hunters. They have curiosity, adventurous spirit. They don’t know where the treasure is, but as they walk, they discover better things.

This doesn’t mean having no goals—rather: The best goals often come from possibilities accidentally discovered in the process, not predetermined at the starting point.

In the circular economy field, what I’m doing now—urban mining digitization, material flow tracking, carbon footprint visualization—no one said ten years ago “this will change the world.” These are all opportunities that emerged through interacting with actual cases, failing, correcting, exploring.

My assumption isn’t “I want to change the circular economy.” The assumption is “this field is interesting, let me see what I can discover.” The results far exceed initial imagination.

Uncomfortable Implications

Pushing Stanley’s logic to its conclusion yields a disturbing conclusion:

If you can completely plan your success, then your ambition might be too small.

True greatness often looks like waste. Looks inefficient. Looks like taking detours.

This means if you want to build an organization capable of producing real innovation, you must give people “permission to waste time.” You must reward novelty, not just progress. You must trust stepping stones, even when they seem completely unrelated in the short term.

Google’s 20% time policy (employees can spend 20% of work time on any project) was once ridiculed as waste. But AdSense came from this 20% time (though Gmail is often mistakenly considered a 20% project, creator Paul Buchheit stated it wasn’t). It was like organized Novelty Search.

More radical application: VC investment. The best startup investments might not be teams with clear business plans, but entrepreneurs with “interesting ideas, currently without clear exits.” Because it’s in pursuing the interesting that they discover opportunities that “planning-type” entrepreneurs never see.

Conclusion: Abandon the Map, Bring a Compass

Stanley’s final line is simple: Stop planning what your life should become, start pursuing what makes you feel alive.

This isn’t motivational fluff. This is mathematically grounded advice.

Multiple experiments prove Novelty Search discovers more powerful solutions than Goal Search. Pursuing novelty goes further than straight-line rushing toward goals.

My twenty years of entrepreneurial experience confirms this. Every decision that seemed most like “waste” later became the most valuable asset.

So if you’re now at some crossroads in life—not knowing whether to “stick to the plan” or “pursue interests”:

Please choose interests. Find those stepping stones that seem outside the plan but genuinely make you curious.

Everyone in history who changed the world was walking such paths. Not the shortest path, but the most interesting path.

And interesting is often the only direction toward greatness.