This week, a Juejin case study gave us a specific number: 16,000 particles (a safe sampling upper limit based on image area). One front-end dev, pairing up with AI, used these particles in Canvas (a web drawing tool) to recreate a Logo particle animation that would normally have had to be done by a UI designer in AE (motion-design software designers rely on) — no AE touched, no outsourcing hired. What we should care about is this — the "one person + AI equals one role" story is quietly flattening cross-role boundaries inside SMBs.

What This Is

The brief was to swap the login page's Logo animation for a new Logo. The original animation was built by a UI designer in AE, exported to JSON (the format Lottie needs to render) via the Bodymovin plugin, then rendered by the front-end using Lottie — meaning it could not be edited, only rebuilt. There was no full-time UI designer on staff, and outsourced schedules were uncontrollable. After staring at the animation for a few minutes, this front-end dev realized: the silky "explode → converge" effect was, at its core, just a particle system (many small points moving by mathematical rules to simulate an animation).

Pairing with AI, he started from a single PNG: scan the pixels as particle starting positions, give each particle a "home position" and an "explode position," and use a 0–1 progress value to interpolate between the two. The result was the same effect — and going forward, swapping a Logo means changing one image path, with no AE source file dependency.

Industry View

Supporters frame it as a textbook case of "small-role gap-filling": what used to take a UI designer and a front-end dev coordinating for a week is now done by one person plus AI in a day. SMBs that can't afford a full-time UI designer benefit especially.

Critics push back: particle animation is just a visual surface front-end happens to be able to recreate by coincidence. UI's real value lies in interaction design, brand consistency, and design-system accumulation — areas AI can't yet touch. Uncontrollable outsourcing can be fixed with SLAs (service-level agreements) and project management; you don't need to detour through AI for that.

There's also a risk flag worth raising: the recreation in this piece is built on "we already know what the animation looks like" — it's reverse-engineering, with replication as the goal. Designing a brand-new motion from scratch, AI still cannot propose original work. Using AI to execute work and using AI to create are two different things.

Impact on Regular People

  • For enterprise IT: design assets dependent on AE / Sketch / Figma source files will gradually be replaced by "rewrite it in code." Front-end devs need to start picking up design fundamentals; UI designers need to start picking up code fundamentals.
  • For individual careers: the "small tasks" that cross role boundaries are becoming "general-purpose skills." Anyone living off a single role needs to ready an adjacent skill as backup.
  • For consumer markets: small changes like swapping a Logo or editing a Banner will get faster and cheaper — but overall page "brand feel" may also homogenize as anyone can tweak it.