1. The zones of uncertainty

In Roadside Picnic, the Zones are places where alien artifacts lie scattered, but human science cannot fully explain them. Stalkers enter to retrieve objects whose functions they do not understand. The book is about operating in a space where your map is incomplete and your categories may be wrong.

The desktop programme is not the Zone, but it has its own zones of uncertainty. This entry tries to map them.

2. Known risks versus genuine uncertainty

There is a difference between a known risk and genuine uncertainty:

  • Known risk: the probability distribution is understood, even if the outcome is bad. Attachment wear is a known risk; we can estimate cycle life and test for it.
  • Genuine uncertainty: we do not even know what the relevant variables are. The market for orbital compute in five years is genuine uncertainty; the customers may not exist yet, or they may want something different.

The programme can manage known risks with engineering. It can only manage genuine uncertainty with optionality and adaptation.

3. The uncertainties

The biggest uncertainties right now are:

  • Market timing: will there be customers willing to pay for the desktop’s services when it is ready?
  • Launch cost trajectory: will launch prices continue to fall, or will they plateau?
  • Regulatory environment: will national licensing become more restrictive or more streamlined?
  • Competition: will large constellations or cloud providers make the desktop’s niche irrelevant?
  • Technology surprises: will a new technology change what the desktop should be?

None of these can be resolved by analysis. They can only be watched and responded to.

4. The response to uncertainty

The response is not to predict. It is to keep options open:

  • Design modularity so the desktop can change its payload mix.
  • Maintain multiple launch provider options.
  • Build relationships before contracts are needed.
  • Keep cash and schedule reserves for surprises.

This is the Popperian approach applied to programme management. Conjectures are made, but the programme stays ready to revise them.

5. What this changes

  • Known risks and genuine uncertainty are distinguished.
  • Uncertainties are listed without pretending they can be solved by more analysis.
  • The response is optionality, not prediction.