Chain-Link Illegibility

We’re building infrastructure (SmartBoxes) to enable a defensible product (Murphy) that incumbents can’t copy without breaking their social contract. But the chain only makes sense when both links exist and work together. No amount of explanation makes this legible to outsiders.

The Insight

Systems of record are specialised and siloed. Even with integrations between every platform, that’s n² OAuth flows per user, n² custom agents that can be chained together. Context doesn’t flow. Users drown in fragmented agents and siloed data.

Developers already solved this for themselves. Coding agents with filesystem access, capability-scoped tokens, persistent context across systems. They have cohesive conversations with agents that can operate across systems, writing notes and scripts to aid future requests.

SmartBoxes = that same pattern for non-developers. Personal, organised like repositories, persistent context. The bitter lesson applied to enabling everyone to be aided by AI.

Why Incumbents Can’t Copy This

Anthropic, OpenAI, Google, Microsoft all understand this trajectory. But with 1B+ users, their appetite for autonomy backlash is zero. They’ll inch forward slowly, constrained by their median user’s comfort level.

Google couldn’t sell Gemini without OpenAI first normalising the form through ChatGPT. Now ChatGPT is the incumbent, they avoid backlash too.

Our advantage: We don’t cater to the median. We target early adopters who want more agency than incumbents will ship.

The Chain-Link Structure

SmartBoxes (wedge / infrastructure)
    ↓ enables
Murphy (moat / value capture)
    ↓ protected by
"Too political for big players"

Murphy is where we capture long-term value. It’s an AI-assisted planning engine—a contract with oneself describing future actions and commitments. This is too political to provide on top of the tools used to implement the plan. Planning must be separate, owned by the user.

SmartBoxes is the required architecture for Murphy agents to do their job effectively across systems.

The chain is linked. Murphy without SmartBoxes doesn’t work. SmartBoxes without Murphy is commodity infrastructure.

Why This Is The Critical Challenge

The problem isn’t distribution. It’s that the chain is illegible until demonstrated:

Can’t sell Murphy without SmartBoxes working
Can’t sell SmartBoxes without the Murphy vision
Leading with Murphy is too ambitious
Leading with SmartBoxes is too abstract
Each link is hard to sell standalone

Our strategy only becomes legible after both links exist.

The Path Through

Dogfood until it’s legible.

We are the first customer. We use SmartBoxes for our own work. We use Murphy to plan our own projects. We become the case study.

The validation gate: When a client uses Murphy to agree the roadmap of their project, we have a product. That moment proves:

External user (not us)
Real stakes (actual project)
The Murphy thesis (planning as contract)
Transferability (works for someone who didn’t build it)

Until then, we’re in the proving period.

Current State

Component	Status
Product form clarity	Clear
Client projects	Exist, structured as requirements/issues
Data structure	Ready to map into PERT-like tree
SmartBoxes (ship-box-dev)	~90% — needs debugging/polish
Murphy	Blocked on SmartBoxes
Client validation	Blocked on Murphy

Critical path: Debug ship-box-dev → Build Murphy on it → Present roadmap to client → Validation

Timeline: Ship-box-dev MVP expected in weeks.

What Doesn’t Work

“Build it and they will come” — The chain is illegible; nobody will come until it’s demonstrated
Explaining the vision — Too abstract; sounds like hand-waving
Leading with SmartBoxes — Commodity infrastructure positioning
Leading with Murphy — Too ambitious without the infrastructure working
Waiting for distribution — Distribution follows proof, not the other way around

The Breakthrough Required

Ship. Dogfood. Demonstrate. Then the chain becomes legible, and distribution follows proof.