The Tyranny of the Roadmap: Why Greatness Cannot Be Planned

In the previous articles, we designed a machine: A high-frequency, self-service VC model that deploys capital dynamically. But why is this machine better at finding outliers than a room full of smart partners?

The answer lies in a counter-intuitive discovery from the field of Artificial Intelligence, detailed in the seminal book "Why Greatness Cannot Be Planned" by Kenneth Stanley and Joel Lehman.

The central thesis is terrifying to the corporate mind: Setting a specific, ambitious objective actually prevents you from achieving it.

In Venture Capital, the entire industry is built on "The Objective."

• "Show me your 5-year roadmap."

• "What is your path to $100M ARR?"

• "How do you capture 10% of the TAM?"

We demand that founders act like archers aiming at a bullseye. But Stanley’s research proves that innovation doesn't work like archery. It works like Evolution. And evolution has no objective.

The Picbreeder Paradox

Stanley’s insight came from an experiment called Picbreeder. It was a platform where users could "breed" images. You started with a blob of noise, and you selected the "child" images that looked interesting.

Here is the kicker: Users who tried to breed a picture of a butterfly failed. They made ugly, distorted blobs. The user who actually bred a perfect butterfly wasn't trying to make one. They were just selecting images that looked "sort of curvy" or "interesting." They followed a trail of curiosity, and suddenly—butterfly. Stanley calls this "The Myth of the Objective."

If you fixate on the destination (The Butterfly / The $10B IPO), you will ignore the Stepping Stones required to get there, because those stepping stones often look like failures.

The "Stepping Stone" Problem in VC

The history of the biggest Venture Capital wins confirms this theory. They almost never started with the objective they achieved.

• YouTube was a video dating site (Tune In, Hook Up).

• Slack was a failed video game (Glitch).

• Twitter was a side-project for a podcasting company (Odeo).

• OpenAI started as a non-profit to "check" AI safety, not a product company.

If a VC had asked the Flickr founders (who were building a game called Game Neverending), "Show me the roadmap to becoming the world's biggest photo storage site," the founders would have failed. There was no roadmap. The photo-sharing feature was just a "Stepping Stone"—a weird little tool inside the game.

Traditional VC kills stepping stones.

When a partner demands a clear path to a massive TAM, they force the founder to optimize for the known rather than explore the unknown. They force the founder to walk in a straight line. But in the landscape of innovation, the path to the highest peak is almost never a straight line. It is a winding, confusing path through the "Valley of Weird."

Novelty Search vs. Objective Search

Stanley introduces the concept of Novelty Search. Instead of optimizing for "Success" (which is deceptive), you optimize for "Novelty" (is this new and interesting?).

This is the scientific argument for our Self-Service / Tier 0 model.

A traditional VC practices Objective Search. They filter for companies that look like "Winners" (Stanford grads, hot sector, clear TAM). This creates convergence. Everyone funds the same "safe" ideas.

Our model practices Novelty Search.

• Tier 0/1 does not ask: "Will this be a unicorn?" (We can't know).

• It asks: "Is this interesting? Is the usage data weird? Is the code velocity high?"

By deploying small checks ($10k) into thousands of "Stepping Stones," we buy the option to see where the path leads. We fund the dating site (YouTube) because the user data is interesting, not because the business plan is solid.

The Deception of the Local Optima

In AI, if you simply try to "climb the hill" (optimize for immediate revenue/growth), you get stuck on a "Local Optima"—a small hill that feels high until you realize you are miles away from Mount Everest.

To find Mount Everest, you often have to go down the small hill and cross a valley of no revenue.

Human VCs hate going down the hill. They panic when a pivot happens. They pull funding when the metrics get "messy." Algorithmically driven VC (Permanent Capital) is patient. We understand that the dip is often the bridge to the higher peak.

The Treasure Hunter vs. The Architect

The current VC industry is filled with Architects. They want to see the blueprints before they lay a brick. They believe greatness can be engineered.

But true innovation is a Treasure Hunt. You don't need a blueprint; you need a metal detector.

• The Blueprint is the Pitch Deck (The Shell).

• The Metal Detector is the Venture Operating System (The Ghost).

We cannot plan the next Google. No one can. But we can build a system that maximizes our exposure to "stepping stones"—funding the weird, the novel, and the interesting—and following the evolutionary path wherever it leads.

As Kenneth Stanley proved: To achieve the highest objective, you must be willing to abandon the objective.

Learn more:

---

Learn more about Meritocratic.Capital