2.4 million people saw this tweet. Design Twitter spiraled. Product people got defensive. Founders got excited.
But the real question is simple: Does an AI-native world make UI less important, or more important than ever?
The answer is more complicated than either camp wants to admit. Some products will shift toward invisible, AI-driven experiences. Others will double down on high-craft UI as their main differentiator.
To understand why, you need to look at what's really underneath this debate: Control.
How AI changes what UI actually does
Much of today’s UI was built for a world where software sits still until you click something.
But the emerging pattern, echoed by several builders on Twitter, is that software is starting to react to what you intend, not what you tap.
Anton Maghami summarized it well:
“Pre-AI was about navigation. Post-AI is about intention.”
The clearest example came from Kousuke A., describing what a truly AI-native to-do list would look like:
“A real AI to-do app wouldn’t ask you to add tasks. It would find them — in your chats, emails, calendar, screenshots, notes, basically your life. Your job becomes to approve or reject.”
This isn't speculation. Products are already moving in this direction.
The UI shifts from managing tasks manually to expressing intention:
”Light tasks today”
”Clear the emails in my backlog”
”Only essential meetings this week”
The AI handles organizing, prioritizing, and scheduling. You become the orchestrator, not the operator.
This is the logic behind Naval's claim: if AI handles the workflow, traditional UI becomes unnecessary.
He's not wrong. But he's not fully right either.
For every ”UI is dead” claim, there's an equally strong rebuttal.
Brad Kowalko said it best:
“UI inspires, educates, guides, conveys information, confirms, delights. The future is a mix of UI and natural language.”
People don't want to just talk to software. They want to see, feel, manipulate, and verify what's happening, especially when AI makes decisions on their behalf.
“Creating through management alone is like painting with boxing gloves on. Great creators still need to tweak and get their hands dirty.”
Chat isn't enough. For some categories, it never will be.
Where “UI matters less in the age of AI” is actually true
There are categories where UI genuinely does fade into the background:
Social platforms: Facebook, Instagram, TikTok; their differentiation isn’t the UI.
It’s the algorithm. Users stay because the feed understands them, not because the interface is beautiful. AI already dictates behaviour here, and large datasets give it the perfect foundation to go fully intention-driven.
E-commerce: Amazon has A/B-tested its way into UI sameness. Most e-commerce platforms now converge on identical patterns, not out of design laziness, but statistical optimization.
UI’s job is not to delight. It’s to convert. These are systems where AI is the product, and UI is just a display surface.
In these categories, Naval is right: UI really is pre-AI.
📰
Join 11,253+ Designers for FREE weekly UX Insights
Every Wednesday, I send out 1 actionable framework to grow your UX career 🌱 — No fluff. Always 2 minutes or less.
Where “UI still matters in the age of AI” is actually true
Then you have the opposite category: tools whose entire differentiation is high-craft, opinionated UI.
Linear: At its core, it’s an issue tracker. But people worship it for speed, taste, restraint, clarity. The UI is the product.
Basecamp & mymind: Their features aren’t exotic. Their worldview is. They win not by being AI-native, but by being design-native. Their users choose them because they believe in the tool’s philosophy, one expressed through interface choices.
In these cases, UI isn’t pre-AI. UI is post-AI.
A piano doesn’t stop being a perfect interface just because music software exists.
Post-AI: AI → API → Data (Humans declare intentions)
This means…
Users speak intentions. Agents operate across tools. APIs become the real touchpoints. Workflows run in the background without anyone having to tap anything.
Software stops waiting for commands. It starts anticipating them.
What UI designers should do in an AI world
The future won't have no UI. It will have more polarized UI.
On one end: invisible UX. Background agents, predicted actions, self-updating states, workflows that execute without intervention. UI becomes thinner because the decision-making is thicker. You won't need screens to move work forward, only to confirm what's happening.
On the other end: high-craft interfaces that feel more intentional than anything we've built in the last decade. These will be the tools people choose rather than tolerate. Tools where design becomes the brand, the moat, and the worldview. These interfaces will matter more because AI will commoditize everything else.
Between them sits a new layer: intelligence design.
What UI designers should do in an AI world
Designers won't just arrange screens. They'll shape how systems behave, predict, respond, and self-correct. They'll train models on user intent, define thresholds for autonomy, decide when to surface decisions and when to execute them quietly.
This is UI beyond layout. This is UI as systems behaviour.
And we'll need a new control layer: UI that confirms, calibrates, and corrects AI when it gets things wrong. These surfaces will be simple but ethically important; the moments where humans reclaim control from automated systems.
Design isn't dying. It's expanding upward.
So… is UI pre-AI? Or post-AI?
AI will erase UI where optimization rules. AI will make UI more important where worldview rules.
Your job as a designer is to recognise the difference.
The future won't be built by people who worship screens or by people who reject them. It will be built by people who understand when to make UI vanish and when to make it unforgettable.
👉
Whenever you're ready, there are 4 ways I can help you: