What Anime Knows About UX That Most Product Teams Don't

Most product design conversations start with wireframes, components, and user flows. Rarely do they start with: "What should this moment feel like?" That's a mistake — and anime might be the unlikely teacher we need.

Your product has a rhythm problem

Alan Cohen published a piece this week comparing anime storytelling with Western superhero films, and it makes an argument I haven't seen made this clearly before. There are fundamentally different models for handling emotion in a narrative — and they map directly onto product design.

Anime (he uses Dan Da Dan as an example) tends toward what he calls "Emotion in Flow." Moods shift gradually, building and releasing tension as part of a continuous arc. Marvel and DC films lean on "Emotion in Conflict" — sharp tonal breaks, dramatic reversals, big moments that interrupt the flow.

Neither is wrong. But here's the thing: most digital products default to conflict mode. Error states that shout. Success screens that celebrate. Loading states that just... exist. We rarely think about the pacing between these moments.

How does a user's emotional state move through a checkout flow? What does the transition between browsing and committing actually feel like? Cohen's framework gives designers a vocabulary for this. And it's the kind of thinking no amount of AI tooling will generate for you.

If you do one thing differently next week, try mapping the emotional arc of a flow you're working on. Not the user journey — the feeling journey. Where does tension build? Where does it release? Is the pacing deliberate or accidental?

"Why make game art at all?"

Speaking of things AI can't decide for you — NVIDIA's DLSS 5 reveal sparked a sharp backlash this week. Dave Oshry, CEO of New Blood Interactive and the developer behind Dusk, put it bluntly: "At this rate why make game art at all?"

The concern isn't really about upscaling. It's about ML-based rendering technology that doesn't just enhance what artists create — it overrides it. When an AI system decides what a frame should look like based on predicted data rather than authored pixels, you've moved from tool to uninvited co-author.

This is a subtler debate than "should you use Midjourney for concept art." It's about whether the final pixel the player sees is the one the artist intended. For studios that treat visual style as a core part of their identity — especially indie studios with distinctive art directions — that's not a trivial question.

Oshry's frustration resonates beyond games, too. Any time a tool quietly substitutes its own judgement for yours, you've lost something. The question designers should be asking isn't "is this technology impressive?" It's "does this technology respect my decisions?"

Two humble UX problems we still haven't solved

Meanwhile, on the less dramatic end of the spectrum, two practical UX stories caught my eye.

First: Vitaly Friedman published a decision tree for choosing between modal dialogs and separate pages. Honestly, it should be pinned to every product team's wall. We all know modals are overused. But "don't use modals" isn't useful advice. The real question is when they're the right pattern and when they're not. Friedman ties it to task complexity and context preservation — which is exactly the right lens.

Second: the "Big Box Paradox." Users leave your site to Google your site. Internal search is almost always worse than external search, and we've somehow accepted that as normal. The case for treating site search as a first-class UX concern, not an infrastructure afterthought, is compelling and long overdue.

Both of these stories are about decisions that require human judgement. No AI is going to tell you whether this particular interaction deserves a modal or a page. No algorithm is going to prioritise your site search because it understands user frustration.

The bigger picture

Here's the thread connecting all of today's stories: the most valuable design work is about judgement, not execution.

Knowing how to pace emotion through a user flow. Deciding whether a modal or a page serves the interaction better. Choosing to prioritise search because you understand how people actually behave. Pushing back on rendering tech that undermines artistic intent.

None of these are automatable. They require taste, context, and a point of view.

That's not an anti-AI argument. AI tools are genuinely useful for execution — generating variations, testing layouts, scaling content. But the decisions about what to build, how it should feel, and what trade-offs to accept? That's still us. And right now, I'd argue it's the part of the job we should be investing in most.

So basically…

The most interesting design thinking this week has nothing to do with new tools or AI features. It's about pacing, emotion, and knowing when to use — or refuse — a particular pattern. Whether you're designing a checkout flow or pushing back on a rendering engine, the skill that matters most is judgement.

And you can't install a plugin for that.