guest@reformer:~$ cat ~/theses/README.md
Theses
What I believe about building software, and why.
Theses on the Craft of Software
27 claims from twelve years of building software. I am writing a short essay for each one. When I learn something new, the list grows.
- Thesis I: Software engineering was never “writing code.” Code was always a means to an end.
#craft · essay coming soon
- Thesis II: The job was always problem solving. Agents just made that obvious.
#craft · essay coming soon
- Thesis III: Seniority is the wisdom you put into a plan.
#craft · essay coming soon
- Thesis IV: An engineer without product sense builds the wrong thing correctly.
#product · essay coming soon
- Thesis V: The most valuable word an architect knows is “no.”
#architecture · essay coming soon
- Thesis VI: Unclear words make coupled code.
#design · essay coming soon
- Thesis VII: Microservices too early and a monolith split too late are the same mistake.
#architecture · essay coming soon
- Thesis VIII: Performance is a design decision, not a tuning step.
#design · essay coming soon
- Thesis IX: Correct first. Everything else second. Elegant and wrong is still wrong.
#craft · essay coming soon
- Thesis XI: Rewrite the small things. Strangle the big ones.
#legacy · essay coming soon
- Thesis XII: The amount of debt decides whether to act. The size of the code decides how.
#legacy · essay coming soon
- Thesis XIII: “We will fix it later” is not on any roadmap.
#legacy · essay coming soon
- Thesis XIV: Vague requirements were always the bug. Agents just ship it faster.
#specs · essay coming soon
- Thesis XV: Agents do not invent bad code. They faithfully build vague requirements.
#specs · essay coming soon
- Thesis XVI: Spend your best model on research and your best attention on the plan.
#specs · essay coming soon
- Thesis XVII: A cheap model can implement a good plan. No model can implement a bad one.
#specs · essay coming soon
- Thesis XVIII: A smaller context makes a smarter agent.
#agents · essay coming soon
- Thesis XIX: Put a gateway between your product and every model.
#ai · essay coming soon
- Thesis XX: Put AI where a wrong answer is cheap to catch.
#ai · essay coming soon
- Thesis XXI: Do not trust the agent. Trust the evidence.
#verification · essay coming soon
- Thesis XXII: The question was never “who wrote it?” It is “what proves it works?”
#verification · essay coming soon
- Thesis XXIII: An agent must never judge its own permissions. That check lives outside the agent, and it should be deterministic.
#agents · essay coming soon
- Thesis XXIV: Hire for judgment. Syntax is now free.
#teams · essay coming soon
- Thesis XXV: Judgment is learned by watching judgment. Pair juniors with seniors.
#teams · essay coming soon
- Thesis XXVI: Team structure is architecture. Design it with the same care.
#teams · essay coming soon
- Thesis XXVII: Every person you add to a team adds another lossy link.
#teams · essay coming soon