Skip to content
Graph of Plan

Can Ship Fast Enough

A single founder with AI assistance can ship four products fast enough to test them before runway exhausts.

The Assumption

We’re betting that modern AI tools (Claude Code, Cursor, Copilot) provide enough productivity multiplier that a solo founder can:

  • Ship SmartBoxes MVP in 4 months
  • Launch Murphy alpha in parallel
  • Get to revenue before runway exhausts
  • Iterate based on feedback without burning out

This is an operational bet. If shipping takes 2x longer than planned, we won’t have time to iterate and find product-market fit before running out of money.

Evidence

Current evidence:

  • Graph-of-plan shipped in ~2 weeks with AI assistance
  • Shipbox prototype functional with agent-heavy development
  • AI tools measurably improving research and boilerplate velocity

Industry signals:

  • Solo founders shipping faster than ever (visible in Product Hunt, Indie Hackers)
  • AI-assisted development becoming standard practice
  • Successful precedents: Pieter Levels, Marc Lou shipping multiple products solo

Counter-signals:

  • AI assistance has limits—novel architecture still requires deep thinking
  • Integration work, edge cases, production hardening take human judgment
  • Context-switching between products has hidden costs
  • AI-generated code sometimes requires significant cleanup

Counter-Evidence

What would prove this wrong:

  • Shipping takes >3x planned time consistently
  • AI assistance doesn’t improve velocity on complex work
  • Runway exhausts before any product reaches validation
  • Burnout or health issues force slowdown

Warning signs:

  • Milestones consistently slipping
  • More time fixing AI-generated bugs than writing code
  • Quality degrading under time pressure
  • Finding excuses to avoid shipping

Impact If Wrong

Products affected: All products delayed, but especially later-sequenced ones (Murphy, Nomos Cloud, P4gent)

Runway impact:

  • At current burn (~£3K/month), runway exhausts in ~18 months
  • If shipping takes 2x longer, we get one product validated instead of 2-3
  • If 3x longer, we’re dead before meaningful revenue

Strategic impact:

  • May need to reduce scope (fewer products)
  • May need to find co-founder or contractor help
  • May need to raise capital earlier than planned

Testing Plan

Velocity tracking:

  1. Weekly: Hours worked vs. features shipped
  2. Monthly: Milestone progress against plan
  3. Quarterly: Runway recalculation with actual burn

Leading indicators:

  • Are demos happening on schedule?
  • Are features completing in estimated time?
  • Is AI assistance saving time or creating rework?

Course corrections:

  • If >2x behind at month 3: Reduce product scope
  • If >2x behind at month 6: Seek help (co-founder, contractor, capital)
  • If >3x behind: Consider pivot or shutdown

Kill criteria: If Month 4 arrives without functional SmartBoxes MVP, dramatically reassess the plan.

Creates risk:

Supports:

Validated by milestones:

Assumption

A single founder with AI assistance can ship four products fast enough to test them before runway exhausts.

Depends On

This assumption only matters if these are true:

Enables

If this assumption is true, these become relevant:

How To Test

Track actual shipping velocity against plan; measure AI productivity gains; monitor runway burn.

Validation Criteria

This assumption is validated if:

  • MVP shipped within 2x planned timeline
  • AI tools providing measurable productivity boost
  • Runway extends to validation point

Invalidation Criteria

This assumption is invalidated if:

  • Shipping takes over 3x planned time
  • AI assistance doesn’t improve velocity
  • Runway exhausts before validation

Current Evidence

  • Graph-of-plan shipped with AI assistance
  • Shipbox prototype functional

Decisions Depending On This