In December 2023, the European Commission formally blocked Adobe's plan to acquire Figma. The deal was valued at $20 billion — roughly 28 trillion Korean won. The U.S. Department of Justice was preparing a similar antitrust case around the same time, and Adobe ultimately abandoned the acquisition in early 2024. At the time, many observers believed Figma had lost its best shot at an exit. A $20 billion acquisition offer is the kind of opportunity that rarely comes twice for a software startup's founders, and when such a deal collapses due to external forces, it can damage both morale and direction.

Yet Figma didn't falter after the deal fell through. Choosing to go it alone, the company aggressively added AI features, expanded its developer-collaboration tools, and kept growing the platform. In a recent interview, CEO Dylan Field said AI has been "a tailwind, not a headwind" for the company — arguing that what was once seen as its biggest threat has actually become a growth engine.

That remark is worth reading for Korean solo planners and one-person PMs — not just because of what it says about Figma's business outlook, but because the shifts underway in the design-tool market bear directly on how solo practitioners actually work.

Why AI Sparked Talk of 'Figma Is Obsolete'

Since its founding in 2012, Figma has reshaped the standard for design-collaboration software. The idea of real-time, browser-based collaboration was unfamiliar at the time, and it quickly reordered a market that Adobe XD and Sketch had dominated. Millions of designers, developers, and PMs worldwide came to collaborate through Figma, and for a long stretch, few teams ran a product-development cycle without it.

But the landscape began to shift after 2024. As large language models like ChatGPT, Claude, and Gemini became everyday tools, a wave of products connected directly to code generation emerged. Vercel's v0.dev instantly produces React components from natural-language prompts alone, while Bolt.new and Lovable can assemble an entire web app from a few lines of text. Skipping the design step entirely and jumping straight to code spread rapidly, especially among startups and solo developers.

Amid this trend, a question surfaced: do you really need Figma? For larger teams sharing a design system and managing revision history, it still holds up — but for early prototypes or MVPs, more people began to say a single AI coding tool is enough. Among small teams, the case for keeping a paid Figma plan started to look weaker.

Field's Logic — and the Skepticism It Draws

Dylan Field offers a different reading of this trend. His argument: the faster AI tools generate code, the more weight falls on the step that defines what should be built in the first place. The quicker the code appears, the more it matters to nail down, in advance, how the resulting screens should look and flow. Because Figma owns that upstream step, he argues, the wider AI spreads, the more — not less — the tool will be used.

Anyone who has actually used an AI coding tool will find this logic at least somewhat familiar. Describing a screen in natural language and designing it visually produce results of very different precision. Vague instructions produce vague outcomes. Field's key point is that a visualized design blueprint can feed an AI far more accurate input than words alone.

Still, plenty of people disagree. If AI coding tools keep improving, a point may come where natural-language descriptions alone are precise enough to produce polished results. Some tools already convert a single screenshot directly into code. Instead of designing a screen in Figma and then handing it to AI, the workflow could flip — AI drafts the screen first, and the user issues revision instructions from there.

There's a counterargument about collaboration value, too. Figma's signature strength — real-time collaboration — scales in value with team size. In settings where multiple designers work simultaneously and developers pull specs directly from design files, Figma's role is clear. But for a solo planner or a two- or three-person team, that advantage shrinks considerably. Whether Field's "tailwind" is actually felt as a tailwind may depend heavily on the kind of organization you're in.

Design Sense Is Migrating Into the Planner's Job Description

What's useful to take from this debate isn't an answer about which tool to pick. It's the broader shift: as AI lowers the barrier to generating UI, visual-design instinct is moving out of the designer's exclusive domain and into territory planners now have to handle themselves.

In the past, a planner wrote a feature spec and a designer built the screens; the planner's involvement in look and flow was limited. That's changed. PMs now sketch wireframes directly, developers review screen flows, and marketers adjust banner layouts. As AI tools lower the entry barrier, the pool of people who work directly with screens has simply gotten bigger.

There's an implication here that's easy to overlook. Which service earns a user's trust, which screen accelerates a purchase decision, which layout lowers bounce rate — none of that is purely a matter of implementing features. The polish and familiarity a customer feels on first seeing a screen translates directly into business results. It's long been an observation in design circles that, when explaining why one version of a product sells better than another, look and feel tend to drive the purchase decision ahead of the feature list. No matter how fast an AI coding tool churns out screens, judging whether the result actually fits the goal — and directing the fixes — still comes down to human eye.

As tools get faster, the gap between planners who have that eye and those who don't is likely to become only more visible.

Where to Start, Right Now, in Practice

It's worth first checking how much of your work still depends on outside designers. If you're still waiting days for a single screen revision, the tools already exist to close that bottleneck yourself, combining AI tools with basic design software. The goal doesn't have to be a "perfect" screen built in Figma's free tier or a similar tool — a more realistic starting point is building it to a level precise enough to hand off accurately to an AI coding tool.

It also helps to set some rules for when to reach for an AI coding tool versus a design tool. AI coding tools turn ideas into screens fast, but they have limits when it comes to fine-tuning the details of a user flow. Design tools take longer but let you hand-adjust layout and visual hierarchy. Lean on only one, and you either gain speed and lose precision, or the reverse.

Working directly with screens builds a vocabulary — for questions like why a layout feels off, or why a color combination fails to build trust. That vocabulary is what lets you give an AI tool precise revision instructions, and it's what lets you give designers more concrete feedback when you collaborate with them. I'd argue this is, right now, the highest-coverage skill a planner can build for the smallest investment. The faster AI tools get, the more the practical value of the judgment that filters their output rises in step.

Just as Figma carved its own path after its $20 billion buyout fell through, solo planners today stand at a similar fork. You can lean on outsourced design and watch the tool paradigm shift around you, or you can walk through the door AI has opened and widen the range of what you can handle yourself. Field's claim that AI is a "tailwind" may not apply to Figma alone. For planners who understand screens, that same wind is blowing in the same direction.