Forget 10,000 hours. Run 10,000 experiments.

There’s one thing that doesn’t get talked about enough when career ladders break, like they have with design.

Classic frameworks, like 10,000 hours, don’t make sense anymore. When the next step isn’t clear, hours aren’t as important as experiments.

So let me run contrary for a second and pick a fight with Malcolm Gladwell.

In case you haven’t heard of it, the 10,000-hour rule is basically about practice.

The world’s best designers, basketball players, violinists, surgeons: they gain mastery in their craft after putting 10,000 hours of practice.

It’s a decent rule of thumb, albeit one that’s been debunked recently.

But it just doesn’t make sense for designers for one key reason. It just assumes the game you’re practicing will still be there when you’re done.

You could have been a polished master of the craft with 10,000 hours on a mechanical typewriter, right as the PC came about.

Or 10,000 hours of being a great event coordinator, and then COVID hits.

Or, just recently, 10,000 hours of design, and then AI changes the game.

10,000 hours only make sense when there’s a clear path forward towards success. You start as a mailroom clerk and you work your way up all the way to CEO.

Every hour compounds because the ladder stays where it is. But that’s no longer the case.

James Altucher has a slightly different number: 10,000 experiments.

Altucher hates waiting in line. That’s why he wrote a book called Skip the Line. In it, he proposes another idea: if you don’t want to wait for permission, run experiments instead.

That mindset is one of the most important things designers need to have nowadays, because the game is changing rapidly.

In Figma's 2026 AI survey, nearly three out of four hiring design managers said they see an increasing need for candidates who are proficient with AI tools.

That seems unthinkable when 2 years ago, at Config 2024, the design community was up in arms about an AI design feature.

When things are moving that fast, 10,000 hours doesn’t make sense.

By the time you get 10,000 hours in ChatGPT, there’s something that blows it out of the water.

You get really good at Replit, and hey, all of a sudden Replit’s not that popular; actually, it’s v0 or Lovable or Claude and MCP or whatever ships next quarter.

Investing 10,000 hours in “AI tools” just doesn’t make any sense, because the tool isn’t the skill.

What matters more than ever is transferability. If you know how to make a NotebookLM second brain, you can probably make a Claude second brain or a Notion second brain.

If that’s the case, experiments, not hours in a single tool, makes the most sense.

For design experiments, it really just consists of three parts:

• What the product was about

• What you asked AI to do, and

• What you did.

Each of these is important, but the 2nd and 3rd are what employers want to see.

Because while some people might work on AI features, most aren’t. Most are working on non-AI projects, but you’ve integrated AI into your workflow.

Maybe you use Figma’s AI to summarize user research.

Maybe you use it to generate first ideas before you start polishing options in Figma.

Or maybe you vibe-code a PR so something actually ships instead of sitting in a spec doc.

AI is somewhere in the project. But what’s the scope? Why did you ask it to do task A but not task B?

How did you set up guardrails and rationale to limit AI to doing this because of X, Y, and Z?

You might not have a clearly defined scope, but saying “Oh yeah, I used AI for this project?” won’t fly.

This is where making it an experiment helps. Saying “I tried to have AI vibe code this, to make a easier PR for engineers to review?” or something similar helps people understand where you’re coming from.

Running experiments is, in a lot of ways, about confronting your discomfort.