🗺️ Roadmap

A roadmap assumes you know where you're going.

The linear model

Quarterly business reviews. Slide decks. Gantt charts. OKRs. The entire apparatus of corporate planning is built on a premise: that you can predict what's valuable before you've built it.

The roadmap says: here's where we are, here's where we're going, here's the timeline. Deviation is failure. Surprise is risk. Discovery is scope creep.

Every feature is rationalized ahead of time. Justified in a meeting. Estimated in story points. Assigned to a sprint. Tracked on a board. The plan is the product. The code is paperwork.

The mycelial model

A slime mold — Physarum polycephalum — has no brain, no plan, no roadmap. It grows toward nutrients. When it finds food, it strengthens the path. When a path leads nowhere, it withdraws. No retrospective needed.

In 2010, researchers put food sources on a map of Tokyo at the locations of major rail stations. The slime mold grew a network that was nearly identical to the actual Tokyo rail system — a system that took decades of central planning to design. The slime mold did it overnight, with no engineer, no budget, no committee.

Same algorithm, different scale: the large-scale structure of the universe — galaxy filaments, voids, clusters — maps onto slime mold networks. Matter flows toward matter along filaments, the same way the slime flows toward food. The cosmos doesn't have a roadmap either.

The difference

The roadmap is a prediction. The mycelium is an exploration.

The roadmap decides what's valuable, then builds it. The mycelium builds toward whatever's nearby, and value emerges from the connections.

The roadmap is afraid of waste. Every feature must justify its existence before it exists. The mycelium wastes constantly — tendrils that go nowhere, paths that dissolve. But the network that remains is optimal, because it was tested by reality, not approved by a committee.

What this means for software

The best features I've shipped were never on a roadmap. They emerged from building something adjacent and noticing a connection. The tunnel reconciler came from needing routes. The fourth architectural principle came from watching an end-to-end test work. A blog post came from a conversation about bug reports.

None of it was planned. All of it was discovered.

The roadmap says "we'll build the email client in Q3." The mycelium says "we built an inbox, then a mailer, then a streaming server, then a triage daemon, and now an email client is just the obvious next step." The destination is the same. The mycelium got there by growing. The roadmap got there by guessing.

Of slide decks and slime molds

A slide deck is a roadmap compressed into 20 minutes. It optimizes for persuasion, not truth. The audience has to believe the prediction before the prediction is tested. The entire incentive structure rewards confidence over exploration.

Nobody ever got funded with a slide that said "we'll grow toward nutrients and see what happens." But that's how every interesting thing actually gets built.

The cult of the pitch

Notice the inversion: the audience has to believe the prediction before the prediction is tested. Belief before evidence, when the actual process is evidence then belief. The system selects for people who can simulate certainty, not people who can navigate uncertainty.

Venture funding is a faith-based process, predicated on the personalities of the leadership. The pitch is the sermon. The deck is the scripture. The partners are the congregation deciding if they believe. The metrics come later to confirm the faith or excuse the heresy.

The result is predictable: overconfidence, brittle plans, post-hoc rationalization. The company that raised $50M can't pivot because the deck said they wouldn't need to. The founder who got funded isn't the one who can find the path — it's the one who can pretend they already know it.

The slime mold doesn't need faith. It has chemistry. The signal is real or it isn't. No personality, no persuasion, no charisma. Just gradients.

Roadmaps are sequential. Mycelium is not.

A roadmap is a line. Point A to point B, with milestones. Progress is measured by distance along the line. If you're not on the line, you're lost.

Mycelium is a network. There is no line. There are tendrils in every direction, strengthening or dissolving based on what they find. Progress isn't distance — it's connectivity. The more connections, the more resilient, the more capable of finding the next nutrient source before you knew it was there.

The roadmap says "don't waste effort." The mycelium says waste is the search space. Every tendril that dissolves was a probe into the wasteland — the uncolonized portion of the problem. The network didn't fail. It mapped the territory. The paths that survive are optimal precisely because the ones that didn't were tried.

Carry me, like you are my brother

Nobody told the slime mold where Tokyo station was. It found it anyway.

Where we're going, we don't need roadmaps.