Medieval warfare optimized around skill that took a lifetime to acquire. The longbow is the cleanest example – not just a weapon but a way of growing up, with strength, technique, and instinct baked in over decades. Then in the 15th century the arquebus arrived: crude, unreliable, and on a per-soldier basis vastly outclassed by a skilled archer or knight, but a sign of an inevitable future.
Guns didn’t demand a lifetime of training or the same physical gifts, and while a state can’t manufacture longbow-men quickly, it can manufacture gunmen. But you couldn’t hand a peasant an arquebus and expect him to fight like a knight – the weapon forced new systems of drill, volley fire, mixed pike-and-shot formations, and powder logistics – and over 300 years those organizational consequences, bigger than the technology itself, produced professional militaries, standing armies, and expanding state capacity.
Arming the engineers
AI coding tools will force a similar shift upon tech companies.
Artemis is already organized around that premise. Our engineering organization is intentionally legible to both humans and machines: systems with clear interfaces, infrastructure expressed in code, fast feedback loops, and a codebase that can be understood and changed without relying on tribal knowledge. That structure is what allows modern tools to do real work. A new engineer can become effective quickly, an AI system can operate with context and guardrails, and improvements in tooling compound across the whole company rather than accumulating around a handful of indispensable people.
Agency is the moat
Organizations that structure themselves to leverage AI will overwhelm incumbents that resist change. This new breed of tech company will onboard engineers faster, iterate faster, and grow faster.
Nobody knows exactly what these structural changes will be. We can guess – low levels of abstraction, everything as code, and extreme testability seem like surefire ingredients in an AI-native codebase. Engineers will remain crucial, but their roles will change. High agency engineers will become increasingly effective as their tools become more powerful. Top performers will function less like elite soldiers and more like generals: defining abstractions, setting constraints, and deciding where risk is acceptable.
AI cannot be an add-on
History’s losers aren’t the ones who dismiss new weapons as inferior; they’re the ones who try to preserve old structures while adopting them. AI won’t fail because it’s worse than senior engineers at writing code. It will succeed because it changes the physics of software development. Companies that cling to hero-driven development will be outflanked by those that rebuild for leverage, not mastery. You can’t bolt firearms onto a medieval army – and you can’t bolt AI onto a 2015 engineering org.
Artemis was built with AI development in mind from day 1. From the org chart to the office layout to the system architecture, we designed Artemis to maximally leverage improvements in foundation model capabilities. These investments have enabled us to deliver best-in-class protection and support to our fast growing customer base from day 1.
