The Platform of Theseus
Issue 237: Transformation, plank-by-plank
The Ship of Theseus Paradox questions whether an object that has had all its components replaced remains fundamentally the same object. Imagine a ship where every plank, nail, and sail is gradually replaced over time—once every part is new, is it the same ship? Further complicating the paradox, if someone were to collect all the original parts and rebuild the ship, which one is the “real” Ship of Theseus? This thought experiment explores the nature of identity, continuity, and change—and applies to everything from personal identity to modern questions about AI and consciousness.
I use the Ship of Theseus as inspiration for design infrastructure teams, who are responsible for maintaining platform elements such as developer experience, infrastructure & reliability, performance, standardization, and scalability. Design infrastructure teams are often responsible for paying off technical debt; usually sins passed to them from previous people on the team. Working on an Angular to React migration project is one of the many rites of passage in working on design infrastructure.
Let’s look at different strategies you can explore to apply the Ship of Theseus method to your migration and re-design. There have been iconic stories of full re-designs that worked. These stories often occur in the early stages of companies. In reality, many designers inherit tech and organizational debt. This is where applying the Ship of Theseus to your platform strategy can help.
Building new planks
As you build new planks to replace, introduce new capabilities and advancements in each one. In any migration project, strive for where you want to leapfrog the platform. It’s a miss to migrate the old components to the new ones.
When you have the opportunity in a migration project to build a new plank, make sure you seize the opportunity. At the surface, the component may look identical to the previous, but there may be new underlying capabilities you want to make sure you ship. This might be improving accessibility features or the extensibility of the component.
Ship small pieces to production
All great design systems started as a button. It’s not because the button has the most tokens and components to test. It’s because a button is small enough to get new components into production. Shipping the tiniest component allows it to adapt to the new ecosystem for stress testing. For example, if build a new component that’s AI-native or has new performance logic to it, testing it internally first through dogfooding allows a deeper understanding of the new element in the ecosystem.
Know when to incur debt
Being on a platform or infrastructure team is a constant tension between scaling architectural excellence and making a trade-off on the speed of delivery. Naturally when you work on a platform or infrastructure team, you want to build things, “the right way.” I’ve been a proponent of grounding your decisions in the reality of the business and customer needs. If that means you have to incur debt that you’ll pay off in the future to make a partnership happen, then it’s worth the debt.
Always be re-designing
Many have been in a situation where they re-design a product and hope that someday it’ll come to fruition. Have the perspective of constantly re-designing as you ship work. It is indeed important to have a north-star concept that guides the work. However, I’ve found it more productive to start building the planks and see how they are functioning in the new system. You don’t have to ship to production each time and can feature flag it until it meets the quality bar you seek.
Design infrastructure teams benefit from having Design Engineers and UI Prototypers who can stress test the North Star. If you constantly have designers experimenting with new patterns that can be integrated in the future, the weight of it is more manageable that’s complete re-design of a legacy system, which could take years.
Ship your Ship of Theseus
I hope this gives you a few thoughts on how to have a future forward platform strategy. A migration or tech debt project is an opportunity to re-imagine the system and the product. In the world of AI-native products, you may find the ship looks different with new planks, but yet holds the essence of what made it special.
Hyperlinks + notes
A collection of references for this post, updates, and weekly reads.
Christina White and Jenny Wen are speaking at Hatch Conference this year. Apply for invite
Apple’s Siri Chief Calls AI Delays Ugly and Embarrassing, Promises Fixes