JHDD Web Design Report — 2026.09.12

BL/S® transformed names into dynamic 3D structures for the Three.js Conference in Paris, demonstrating a personalized approach to web graphics.

The collective push for bespoke, generative visual systems is evident across these projects, shifting web graphics from static assets or simple interactions to complex, real-time computational art. WebGPU, Three.js, and TSL are not merely rendering tools, but enablers for procedural craft, allowing designers to define visual logic rather than just arranging elements.

The BL/S® experiment of turning names into digital architecture moves beyond traditional web typography into dynamic, personalized layout generation. This is not simply about rendering 3D objects, but about creating responsive, generative visual systems where the very structure and form are dictated by data and authored logic. While many within the industry advocate for minimal, framework-driven designs, often dismissing advanced GPU techniques as “mere eye candy” that inherently hinders performance, this perspective overlooks the fundamental shift occurring. Projects like Still and the “Infinite Liquid Glass Grid” showcase an intent to integrate deeply custom visual computing directly into the browser, delivering unique aesthetic experiences that are impossible with standard CSS or even conventional JavaScript animations alone. The capability to build a living generative garden with procedural systems or refractive liquid grids using TSL points to an emerging discipline where art direction involves authoring performant GPU-driven visual languages. This is about deep craft. Such experiences, when meticulously optimized, leverage WebGPU’s direct hardware access for efficiency, providing rich visual fidelity without necessarily compromising loading times or interaction fluidity. They offer a path to profoundly differentiate a digital presence in a visually homogenous landscape, moving beyond template-driven aesthetics. Within two years, this shift will solidify, with discerning brands increasingly seeking out designers capable of building performant, unique visual systems using WebGPU and TSL for core interface elements, rather than as isolated decorative hero sections. These systems will begin to influence how interactive layouts are conceived, offering a bespoke approach to web typography and responsive visual composition.

The primary resistance to this evolution comes from the prevailing preference for quick development cycles and the reliance on off-the-shelf component libraries, which often prioritize efficiency of assembly over uniqueness of visual experience and deep aesthetic craft. The perceived steep learning curve for mastering technologies like WebGPU and TSL deters many front-end developers, maintaining a significant gap between advanced graphics programmers and those primarily focused on the document object model. This inertia is further reinforced by the current limitations of many content management systems and design tools, which are not yet equipped to seamlessly integrate deeply custom, performant WebGPU experiences directly into their standard workflows, making adoption for mainstream projects seem daunting. The prevailing metrics for web performance also sometimes struggle to account for the qualitative gains in bespoke visual experiences versus the quantitative overhead of advanced rendering.

A working web design professional should dedicate time this week to understanding the core principles of TSL (Three.js Shaders Language) and WebGPU’s capabilities. Experiment with simple shader examples to grasp how a GPU processes visual information, moving beyond declarative CSS properties to procedural visual generation. This practical step, even a small one, initiates the transition from consuming pre-built components to authoring bespoke visual logic.

---

Curated References