JHDD Web Design Report — 2026.09.13
BL/S® demonstrated at the Three.js Conference in Paris that personal names can be transformed into dynamic 3D architectural forms.
This capability represents a pervasive shift in web development, where real-time 3D graphics and complex interactive experiences are becoming central to content delivery, moving beyond decorative elements. The increasing adoption of WebGPU, often paired with Three.js, facilitates a new generation of immersive digital environments that challenge traditional two-dimensional web layouts. This trend connects diverse projects, from Yestalgia’s playful ’90s-inspired interactions for Decathlon to the generative gardens of Still, all leveraging advanced rendering pipelines to create unique user engagement.
The conventional industry opinion often views such intensive 3D applications as niche experiments or significant performance liabilities, unsuitable for mainstream web experiences. This perspective frequently dismisses complex visuals as mere “eye candy” that inherently compromise accessibility and speed. However, this assessment largely overlooks the fundamental advancements WebGPU brings, particularly its direct access to GPU hardware, which fundamentally reshapes what is performant on the web. Projects like the infinite liquid glass grid, combining Three.js, WebGPU, and TSL, illustrate that highly sophisticated visual effects can now operate with unprecedented efficiency, provided they are developed with meticulous front-end craft. The challenge is no longer primarily technical feasibility, but rather the craft of optimizing these rich experiences to uphold core web principles.
BL/S®’s interactive name transformer exemplifies this shift, showcasing generative content that is not merely rendered but dynamically constructed based on user input. This moves beyond pre-rendered animations or static 3D models into truly interactive data visualization and content creation within the browser. Such applications necessitate a deep understanding of performance optimization at the shader level and a re-thinking of how content is authored and delivered. By early 2028, the seamless integration of generative 3D elements, once confined to specialized showcases like Still or BL/S®’s experiment, will be a fundamental expectation for high-end brand presences, moving beyond mere marketing sites into product configurators and educational platforms.
The primary resistance to this evolution comes from the inertia of established tooling ecosystems and the substantial learning curve for front-end development teams. Many dominant frameworks and content management systems prioritize simpler DOM manipulation and CSS-based layouts, abstracting away the low-level rendering control necessary for advanced WebGPU applications. This creates a skills gap and an ecosystem that, while robust for traditional web design, struggles to natively support the advanced real-time rendering paradigms now emerging.
Web design professionals must dedicate focused time to understanding shader languages, perhaps beginning with TSL, and exploring how procedural generation can extend beyond artistic showcases into functional interface design. Experimentation with WebGPU’s direct capabilities, even through simplified examples, will be crucial for adapting to the new demands of visual and interactive craft.
TL;DR
The web is evolving towards highly interactive, 3D-first experiences powered by WebGPU and Three.js, demanding a deeper craft from front-end designers.
Curated References
About this editorial — This piece was developed using AI-assisted research and curation across multiple industry sources. All analysis, opinions, and predictions represent the editorial perspective of JHDD. Sources are linked in the references section above.