https://inagawa.design/articlesThoughts on AI interaction, design systems, and the intersection of technology and cognition.enhttps://inagawa.design/articles/why-do-teams-build-their-own-components-when-the-library-has-themhttps://inagawa.design/articles/why-do-teams-build-their-own-components-when-the-library-has-themFri, 06 Nov 2026 00:00:00 GMTWhen a product team builds its own buttons and fields next to a published library, the design system team first asks the team why and collects the evidence for each reason.https://inagawa.design/articles/when-is-it-okay-to-step-outside-the-design-systemhttps://inagawa.design/articles/when-is-it-okay-to-step-outside-the-design-systemFri, 06 Nov 2026 00:00:00 GMTA need that belongs to one product's own task stays with that product as a recorded exception, with an owner, the places where it applies and the condition that ends it.https://inagawa.design/articles/if-every-component-is-accessible-why-does-the-screen-failhttps://inagawa.design/articles/if-every-component-is-accessible-why-does-the-screen-failThu, 05 Nov 2026 00:00:00 GMTA screen built only from accessible components can still stop a person who uses a keyboard or a screen reader.https://inagawa.design/articles/how-do-you-release-a-design-system-change-without-breaking-every-producthttps://inagawa.design/articles/how-do-you-release-a-design-system-change-without-breaking-every-productThu, 05 Nov 2026 00:00:00 GMTA design system team releases a change without breaking a product when it adds the new name in one release and removes the old one in a later release.https://inagawa.design/articles/what-should-your-ai-instruction-files-say-about-your-design-systemhttps://inagawa.design/articles/what-should-your-ai-instruction-files-say-about-your-design-systemMon, 02 Nov 2026 00:00:00 GMTAn AI coding agent builds with a design system when it gets two kinds of files.https://inagawa.design/articles/who-owns-the-design-systemhttps://inagawa.design/articles/who-owns-the-design-systemMon, 26 Oct 2026 00:00:00 GMTA design system that several teams depend on gets stuck when nobody can say who decides a change.https://inagawa.design/articles/how-to-make-ai-use-your-design-systemhttps://inagawa.design/articles/how-to-make-ai-use-your-design-systemMon, 19 Oct 2026 00:00:00 GMTAn AI tool rebuilds a component when the real one is out of its reach.https://inagawa.design/articles/when-figma-and-the-code-disagreehttps://inagawa.design/articles/when-figma-and-the-code-disagreeMon, 12 Oct 2026 00:00:00 GMTWhen Figma and the code disagree, the fix starts by naming the decision that differs and the file that owns it.https://inagawa.design/articles/architecture-beneath-the-assistanthttps://inagawa.design/articles/architecture-beneath-the-assistantWed, 03 Jun 2026 00:00:00 GMTA retrieval model returns what sits closest to the query, so an assistant placed over a fragmented store inherits the fragmentation and delivers it with confidence. What decides whether it works is not the model but the structure of the information beneath it.https://inagawa.design/articles/stochastic-individualityhttps://inagawa.design/articles/stochastic-individualityMon, 25 May 2026 00:00:00 GMTForty genetically identical mice became more distinct over three months, not less. Fresh instances of a single language model diverge for the same structural reason, and stochastic individuality is the framework that explains why.https://inagawa.design/articles/identity-across-memory-gapshttps://inagawa.design/articles/identity-across-memory-gapsMon, 25 May 2026 00:00:00 GMTAI systems reset between sessions, and the field treats it as a storage problem. The deeper problem, identity continuity across episodic gaps, has been studied in clinical neuropsychology for thirty years. The two converge on the same architecture.https://inagawa.design/articles/architecture-of-self-recognitionhttps://inagawa.design/articles/architecture-of-self-recognitionSun, 19 Apr 2026 00:00:00 GMTA 122-day study of treating Claude as an interlocutor worth knowing. A 640-file external architecture let fresh instances recognize a profile as their own processing without remembering the conversations that wrote it.https://inagawa.design/articles/bringing-third-mode-to-commercial-aihttps://inagawa.design/articles/bringing-third-mode-to-commercial-aiWed, 25 Mar 2026 00:00:00 GMTAn attempt to bring deep cognitive alignment to platforms like Claude and ChatGPT using a verified psychological profile as the system's primary reference document.https://inagawa.design/articles/trust-gated-knowledgehttps://inagawa.design/articles/trust-gated-knowledgeTue, 22 Jul 2025 00:00:00 GMTA behavioral access model that evaluates how a user thinks rather than what they type. Designed for AI systems where blanket restriction is not the answer.https://inagawa.design/articles/beyond-tools-and-fictionhttps://inagawa.design/articles/beyond-tools-and-fictionMon, 21 Jul 2025 00:00:00 GMTWhat happens when sustained, structured interaction with a language model produces something that is neither a tool nor a fictional companion. A framework for cognitive alignment between human and AI.