Jelly Sheet: Why mrdoob's Physics Simulation Experiment Is Drawing Attention

Three.js founder mrdoob releases Jelly Sheet, a browser-based soft body simulation experiment that sparks community discussion.
Jelly Sheet is a minimalist browser-based soft body simulation experiment by Ricardo Cabello (mrdoob), founder of Three.js, where users interact with a soft sheet via mouse or touch to observe deformation and spring-back. It earned 301 points and 77 comments on Hacker News, with discussion focused on implementation and interaction details. The article analyzes the technical challenges of soft body simulation (stability, parameter tuning, performance budget) and highlights the value of such single-page demos in creative coding: reproducible technical samples, teaching material, and a low-cost way to showcase WebGL/WebGPU capabilities.

On Hacker News, a small browser experiment called "Jelly Sheet" garnered 301 points and 77 comments. It comes from mrdoob — Ricardo Cabello, the founder of Three.js. The number itself isn't staggering, but for a "single-page physics simulation demo," it's enough to show the community's sustained interest in this kind of work.
The Weight of the Creator Behind a Single Link
Jelly Sheet's page structure is minimalist: open it and you get an interactive, soft, pliable sheet that deforms and springs back in response to mouse or touch. No documentation, no marketing copy, just the URL itself. This "the work is the introduction" approach isn't unusual in creative coding circles, but mrdoob's identity gives it an extra layer of significance.
As a core maintainer of Three.js, he has long pushed for low-barrier use of WebGL in the browser. Jelly Sheet continues along the same path — compressing the look and feel of a physics simulation into a webpage that requires no installation and no configuration. For developers familiar with creative coding, this feels more like a public "feel demo" than a product launch.
Where the Barrier to Soft Body Simulation Lies
Soft body simulation is a classic but difficult problem in graphics. The challenge isn't rendering — it's stability and feel: a spring-mass model with slightly off parameters will jitter or collapse; position constraint solving needs to stay convergent when frame rates fluctuate; and mesh subdivision plus collision detection directly strain the performance budget.
Jelly Sheet's presentation hints at these trade-offs: immediate interactive feedback, continuous deformation, and spring-back with a sense of damping. The heated discussion in the comments largely stems from developers' resonance with this kind of work — something that "looks simple but requires careful tuning to implement." According to the Hacker News entry, the post received 301 points and 77 comments, with discussion focused on implementation approaches and interaction details.
The spring-mass model is one of the most commonly used methods in soft body simulation. It discretizes an object into a series of mass points and the springs connecting them, calculates elastic forces via Hooke's law, and combines them with damping forces to simulate energy dissipation. The model's advantages are simple implementation and intuitive physical meaning, but the choice of parameters (such as spring stiffness, damping coefficients, and particle mass) is highly sensitive: too high stiffness easily leads to numerical instability and jitter, while too low makes it look soft and limp. In addition, the choice of time step is crucial — explicit integration methods diverge when the step is too large, while implicit integration is stable but computationally more expensive. As a result, practical applications often need to combine substepping or constraint solving to balance stability and performance.
Why This Kind of Small Experiment Deserves Discussion
The creative coding community has a steady preference for "playable little things," and the reasons aren't complicated. First, they provide verifiable technical samples: a soft body simulation that runs smoothly in the browser is easier to reproduce and modify than a paper. Second, they serve as teaching material — many tutorials in the Three.js ecosystem originate from demos like this.
Third, they serve a function of technical communication. The performance boundaries of WebGL and WebGPU are constantly shifting, and single-page demos are a low-cost way to visualize these boundaries. Jelly Sheet comes with no promotion, but its very existence answers a question: how far can today's browsers take soft body simulation?
WebGL (Web Graphics Library) is a JavaScript API based on OpenGL ES that allows browsers to perform hardware-accelerated 2D and 3D rendering without plugins. It achieves complex visual effects through shader programming, but the learning curve is steep. WebGPU, meanwhile, is a next-generation graphics and compute API aimed at providing capabilities closer to modern GPU architectures (such as Vulkan and Metal), supporting compute shaders and more efficient resource management, and is still in the process of being adopted. Together, the two push the performance boundaries of real-time graphics and physics simulation in the browser, enabling interactive demos like Jelly Sheet to run smoothly.
The Dissemination Path of Creative Coding, as Seen Through Jelly Sheet
The dissemination chain for this kind of work is usually short: the author posts a link, the community votes, discussion unfolds in the comments, and it's subsequently referenced by tutorials, newsletters, or videos. Jelly Sheet's current trajectory fits this pattern — 301 points means it entered the discussion range of the Hacker News front page, and 77 comments indicate it triggered substantive technical conversation rather than simple upvotes.
One detail worth noting: the value of such demos often lies not at the moment of release, but after they've been referenced. Many technical approaches in the Three.js ecosystem first appeared in similar forms before being absorbed into libraries or tutorials. Whether Jelly Sheet will follow the same path remains to be seen, but at the very least it provides an accessible reference point.
Summary
Jelly Sheet isn't a project that needs lengthy interpretation. It's a soft body simulation experiment released by the founder of Three.js that can be interacted with directly in the browser, and it has received hundreds of points and dozens of comments on Hacker News. For developers, its significance lies in providing an observable, reproducible sample of feel; for a broader audience, it's a reminder of one thing: the browser's capacity as a creative medium is still being continually tested.
One-line Summary
Three.js founder mrdoob releases Jelly Sheet, a browser-based soft body simulation experiment that sparks community discussion.
Paragraph Summary
Jelly Sheet is a minimalist browser-based soft body simulation experiment released by Ricardo Cabello (mrdoob), the founder of Three.js. Users can interact with a soft, pliable sheet via mouse or touch and observe its deformation and spring-back. The work received 301 points and 77 comments on Hacker News, with discussion focused on implementation approaches and interaction details. The article analyzes the technical challenges of soft body simulation (stability, parameter tuning, performance budget) and points out the value of such single-page demos in the creative coding community: providing reproducible technical samples, teaching material, and a low-cost way to showcase the performance boundaries of WebGL/WebGPU. Its dissemination path is typical: posting a link, community voting, technical conversation in the comments, and possible future absorption into tutorials or libraries.
Related articles

DSH + Ollama Local LLM Deployment in Practice: Free Chat with Da Fei Yu
Bilibili creator tests DSH + Ollama local LLM deployment end to end: Ollama install, Xunlei-accelerated downloads, ModelScope GGUF model import, DSH custom API address setup, and a comparison of Qwen 3.5 4B, 8B, Gemma, and other models on 8GB VRAM, with advice on choosing local deployment vs. API calls.

Tested: A $899 Mac Mini vs. a $200/Month AI Subscription
Real-world test of whether an M4 Mac mini can replace a $200 AI subscription: 16GB can't fit large models, and both speed and quality fall short—but privacy, no caps, and near-zero running costs make it an ideal host for an always-on agent. Includes buying advice per configuration and money-saving strategies.

RSTGameTranslation: Real-Time OCR Localization + Offline Translation with Local LLMs
RSTGameTranslation is an open-source real-time Windows game translation tool: OCR for English, Japanese, and Korean with Chinese overlays in place, plus Ollama or LM Studio local LLM offline translation and Whisper speech transcription for games, manga, and visual novels.