A Realization About Parenting and Startups

My child recently turned four. Like many families, our dinner table conversations began to drift toward extracurriculars. To be honest, I hadn't given education much serious thought. Yet, I felt an underlying unease with the direction of the discussion. Why were we trying to "optimize" such a young child? (And when did parenting start sounding like a quarterly business review?)

Last week, while traveling, I picked up a book titled Discovered, Not Designed by Sean McClure. That book connected the dots for me, linking my questions about raising a child directly to the challenges of building a business. Funny how a book about complexity theory can suddenly make you rethink both your startup strategy and your four-year-old's afternoon schedule

From Design to Discovery

Consider a deceptively simple question: Which is more complex, riding a bicycle or solving a physics problem?

Intuition might say the physics problem. After all, it looks harder on paper, requires formal education, and makes you feel inadequate when you get it wrong. But from an information processing perspective, riding the bike is the true "Hard Problem". When you ride, your brain unconsciously processes a dynamic system full of massive, real-time, fuzzy variables. A physics problem, by contrast, exists in a tidy world of clear boundaries and self-consistent logic.

This contrast reveals a deeper question: Why does the human brain effortlessly handle the "Hard Problem" that would stump a supercomputer, yet struggle with the "Difficult Problem" of clear logic?

The answer lies in two different creative approaches: Design versus Discovery, or careful construction versus natural emergence.

As AI entrepreneurs, we see this daily. AI tools boost our efficiency in coding and writing. But the true shift is not just in efficiency; it is a fundamental change in how we create. Before AI, every great engineering feat, even a microchip, was a product of "Design". Every module's function was clear, crafted by human intelligence from the ground up. We knew exactly why each transistor was where it was. We had blueprints.

Large Language Models (LLMs) represent a different engineering paradigm. Their power does not come from design but is "discovered" within massive datasets. Even the creators of a trillion-parameter model do not know what every parameter does. (Imagine building a skyscraper where you can't explain what 99% of the bolts are for, but somehow it stands anyway.) Humans merely define the ecosystem (framework and data), and the model "evolves" within it.

This shift from "Design blueprint" to "Discovery emergence" is not just a technical novelty. It offers us a completely new way to solve problems, especially the messy ones that refuse to fit on a spreadsheet.

Two Types of Problems, Two Different Approaches

The world presents us with two categories of challenges.

Difficult Problems

These may be complicated, but their causal chains are clear. You can solve them through logic and decomposition. Chip design is a classic example of "Design Thinking". You break high-level needs down into architectures, modules, and circuits, strictly following physical laws. The process is painstaking, but if you follow all the rules, the result is predictable. You can simulate it. You can test it. And when it works, you know exactly why.

Hard Problems

These have a vast "possibility space". Elements lack clear causal links or may even contradict each other. Traditional design fails here. Face recognition is a perfect example. If you tried to "engineer" it by defining rules (distance between eyes, nose shape, lighting conditions), you would fail spectacularly. Lighting, angles, and expressions create infinite variations. To cover all possible human faces under all possible conditions, you would need more rules than atoms in the universe.

Yet, a human recognizes a familiar face instantly. We do not check rules; we rely on a capability that "emerged" from experience, something we cannot explain even if we tried. Engineering finally cracked this code not by designing more clever rules, but by letting neural networks "discover" the patterns themselves through training.

The Trap of Design Thinking: Everyone Wants a Cheat Code

"Design Thinking" offers a fatal temptation when facing a complex world. It suggests that if we just find the right "walkthrough" or cheat code, we can win. It is the mentality of someone who believes life is a video game with a published strategy guide.

This is the "assembly line" approach to parenting. Parents believe that if they stack enough skills (coding today, math Olympiad tomorrow, piano the next day), they will assemble a "perfect" child. It is like they are following a recipe: two cups of STEM, one tablespoon of classical music, a pinch of Mandarin, bake at 350 degrees for 18 years, and voilà, Harvard material.

Reality often delivers a rude shock. The child hates the piano and is confused by coding. Over-design crushes their curiosity. Life is not an assembly line; it is an improvisational performance. And children, inconveniently, did not read the manual.

An elite problem solver's core value is not a stack of skills but a "Super Black Box" intuition, judgment, and creativity that emerged from complexity, not from a syllabus. Skill stacks are replaceable. You can hire those. That black box capability? That is the thing competitors cannot copy, and recruiters cannot poach.

Embracing Complexity: The Path of Discovery

Solving Hard Problems requires a path opposite to Design. It mimics natural evolution. You apply external pressure (also known as "reality") and let the system find its own way through mutation, feedback, and selection.

This is how we train LLMs. We set the frame, feed the data, and provide feedback ("good answer" versus "bad answer"), but we never manually tune the internal parameters. We do not open the hood and start fiddling with wires. The capability emerges, almost like magic, except it is not magic. It is just complexity at work.

This reveals a counterintuitive secret: Complexity is not a burden; it is an advantage. Hard Problems do not rely on rigorous logical deduction. They rely on two capabilities that sound suspiciously like "street smarts":

Pattern Recognition The ability to compress complexity and abstract from it. It sees through the chaos to find stable commonalities. This is the "What" (what is actually happening here beneath all the noise).

Heuristics Shortcuts for decision-making. These are rules of thumb distilled from experience that help us decide "How" to act when the map is useless, and the compass is broken. It is the operating system of people who have seen things go wrong and learned not to repeat the same mistakes.

When teaching a child to ride a bike, we don't lecture them on the dynamics of angular momentum and gyroscopic stability. (Although that would be hilarious to watch.) We run behind them, hold the seat, and shout, "Keep pedaling!" Through wobbles and corrections, they "discover" balance. Once they have it, they have it for life. This is "discovery" in action, and it works better than any instruction manual ever could.

Characteristics of Discovered Solutions: Fast, Flexible, and Utterly Baffling

Solutions that evolve through discovery share three distinct traits that frustrate traditional engineers:

Speed: Once the connection is made, execution is instant. You recognize a friend in milliseconds without running through a mental checklist of facial features. No spreadsheet required.

Flexibility: These solutions are incredibly adaptive. Like Stephen Curry on the basketball court, the "discovered" capability handles infinite variables in real-time. Every play is different, yet he consistently makes shots that make highlight reels. Try writing a manual for that.

Unexplainability: This is the most maddening part. You cannot write a 10-step guide for it. A master craftsman often cannot explain how they do it; they just have "the feel". Ask them to document their process, and you will get a lot of hemming and hawing about "experience" and "intuition". Not helpful for scaling, but essential for excellence.

In this sense, complexity stops being the enemy and becomes the partner. We do not need to "defeat" every complex problem by breaking it into a million tiny pieces. We can embrace the complexity as it exists in the real world, messy and interconnected and alive.

Should We Throw Design Thinking in the Trash?

So, should we discard Design Thinking entirely and go full "discovery mode" on everything? Tempting, but no.

These two modes are partners, not rivals. Discovery sets the course; Design ensures execution. They are like the "visionary founder" and the "operator COO" of your cognitive toolkit. You need both, or things fall apart in predictably different ways.

Take running a successful restaurant chain:

Discovery Thinking: The founder cares about macro properties like customer satisfaction, brand reputation, and overall "dining atmosphere". They think about the restaurant's place in the market. When profits drop, they might intuitively sense "the service experience is off" and demand changes. But they will not micromanage every server's tray-carrying technique. That would be insane.

Design Thinking: The kitchen manager ensures every dish follows strict procedures. The steak must cook at a specific temperature for a specific time to ensure taste and food safety. They care about ingredients, recipes, and steps, the quantifiable details with clear cause and effect.

What Happens If You Only Have One:

If you only have the founder's "discovery vision" without the manager's rigorous execution, food quality collapses, and your reputation craters. If you only have perfect dishes but no sense of the market or customer needs, you might create a Michelin-quality menu that nobody wants. Either way, you are out of business, just for different reasons.

The ideal state: Use "Discovery Thinking" to set direction. Use "Design Thinking" to execute flawlessly.

Four Principles of Discovery Thinking

Discovery thinking follows four core principles: See from the outside in, embrace trial and error, obsess over feedback, and survive first.

See from the Outside In

Do not obsess over internal mechanics. Focus on the system's macro properties and abstract concepts. Because truth often hides in the things that do not change, the enduring patterns that hold across time and context. High-level abstractions outlast low-level details.

Take "network effects" in business:

High-level truth: "The more nodes a platform connects, the more valuable it becomes to each node". This holds in any era.

Low-level details: In the 19th century, it was telephone networks. In the late 20th century, Windows. In the early 21st century, Facebook. In the future, maybe an AI platform.

An entrepreneur who grasps "network effects" as an unchanging principle can spot real opportunities across different technology waves, instead of getting lost in the ever-shifting product landscape. Spot the lighthouse first; then figure out how to sail there.

Embrace Trial and Error

If we have identified a "Hard Problem" but the path is unknown, the answer is simple: Try things.

"Discovery Thinking" does not believe in ivory tower planning. Instead of spending a year building a "perfect" product, ship a minimum viable product (MVP) fast and throw it into the market to test. Like biological evolution, there is no preset blueprint, just countless random mutations and environmental selection. We actively create those mutations and see which survive. "Fire, then aim" often beats meticulous planning in an uncertain world.

Obsess Over Feedback

Beyond hard metrics like download rates and conversion, "Discovery Thinking" cares about "soft" signals: user complaints on social media, emotions in app store reviews, feelings and atmospheres in interviews.

Great entrepreneurs can sense from users' "feelings" whether the product direction is right or wrong, often earlier than the data shows. It is like having a second radar that picks up tremors before the earthquake hits.

Survive First, Optimize Later

Finally, the most practical principle: Survive, then thrive.

In the early startup days, every decision has one ultimate test: Does it work in reality? Does it keep the company alive?

A feature might be "elegant" technically or "perfect" in design, but if users do not care and it does not drive growth or move toward profitability, it must be cut. In brutal market competition, you only get to iterate and "discover" the real success model if you live long enough. Dying elegantly is far worse than surviving ugly.

Uncertainty: The Designer's Nightmare, the Explorer's Playground

In "Design Thinking" mode, uncertainty is the ultimate enemy. Like a precision machine, any part that deviates from the blueprint or any unplanned vibration might trigger a cascade failure. In this mode, people panic at the slightest course deviation and freeze when information is fuzzy.

But building a startup is not building a machine. It is cultivating a rainforest.

In a rainforest, when a tree falls, it does not bring destruction. It brings sunlight to the ferns and fungi below, creating new life. An unexpected wildfire clears dead brush and lets fire-resistant species take root. The ecosystem becomes more resilient, not more fragile.

This is what "Discovery Thinking" gives us: Anti-fragility.

We stop fearing change. We stop treating surprises as enemies. Instead, we treat them as raw material and signals.

Market ambiguity becomes the entry point to clarify real constraints and value with users. Industry disruption becomes the radar signal for discovering new opportunities and migration paths. Internal plan disruptions become the training ground for the organization to evolve stronger adaptability.

Looking back at my years of ship, I notice an interesting pattern. The decisions that actually worked, the ones that moved the needle, were rarely the ones I felt most confident about at the time.

Instead, they were the small, "casual" experiments. Maybe we were addressing a client's last-minute request. Maybe a team member ran an "out-of-scope" test. Maybe we just wanted to validate a "dumb" idea. Those turned out to work like a charm.

But here is the trap. Once you gain some experience, you start believing you can "design" success. You do big strategic plans, detailed business model canvases, and airtight execution roadmaps. You give yourself a million constraints. Then reality changes, users change, and your understanding changes. That beautiful blueprint either gets shelved or becomes a straitjacket, preventing the team from adapting.

So this time, I want to try a different approach: Lightweight design. Not a massive blueprint, but a compass.

Meta-Design: Build the Compass, Not the Map

"Meta-Design" does not mean we stop thinking. It means we change how we think and plan. The core is choosing the right high-level "Properties" for our endeavor at the very beginning, properties that remain stable even as the details shift. This is not some optional philosophical exercise. It is the foundation of the entire "discovery" process.

Back to the metaphor: Traditional "Design" is like drafting a detailed nautical chart before departure, marking every day's route, every port of call, and expected weather. But the ship is sailing an ocean full of unknown storms and currents. That detailed map becomes obsolete almost as soon as you set sail.

"Meta-Design" acknowledges this reality. It tells us that in an uncertain journey, obsessing over a map that will expire is futile. Better to build a reliable compass.

This compass will not tell you to sail 50 nautical miles northeast on day one and 30 miles south on day two. But it will always point toward where you want to go. When the wind shifts, the currents change, or even the destination coordinates need adjusting, the compass still works. Because it anchors to something more fundamental and enduring: the macro attributes that define what kind of organization you want to become and what kind of value you want to create.

How Meta-Design Helps You Navigate

Choose Your North Star, Not Your Route

Instead of planning "Year 1: Build feature X, Year 2: Enter market Y, Year 3: Achieve Z scale", Meta-Design asks you to define the enduring characteristics your company should embody. What are the unchanging qualities that should guide every decision, even when tactics change?

This might be qualities like "always prioritize depth over breadth" or "compound value, not extract it" or "partner with ecosystems, don't parasitize them". These are not tactical goals. They are the fundamental nature you want your organization to have.

Enable Adaptation Without Losing Direction

With Meta-Design, when reality forces a pivot, you do not feel lost. You ask: Does this new opportunity align with our core attributes? If yes, pursue it boldly. If no, pass, no matter how tempting. The compass keeps pointing north even when the ocean is chaotic.

Create the Frame for Evolution

Like training an LLM, you set the environment and constraints (your meta-attributes), then let the organization experiment, learn, and evolve toward better solutions. You provide feedback on whether actions strengthened or weakened those core attributes. Over time, the right strategy emerges naturally, discovered rather than dictated.

Reduce Analysis Paralysis

When facing complex decisions with incomplete information (which is always), you do not need perfect data. You check: Does this feel aligned with who we are trying to become? That "feel", internalized from your meta-attributes, becomes your decision heuristic. It is faster than analysis and often more accurate in uncertainty.

The beauty of Meta-Design is that it liberates you from the tyranny of the perfect plan. You are no longer trying to predict an unpredictable future. You are building a navigation system that works regardless of which storms you encounter. You focus your planning energy on the attributes that compound value over time, not the tactics that expire next quarter.

In a world of relentless change, the compass matters far more than the map.

Conclusion

• Use Discovery Thinking to set direction and navigate uncertainty
• Use Design Thinking to execute with rigor once the path is clear
• In a world where change is constant, your compass (meta-attributes) matters infinitely more than your map (detailed plans)