Every time tools get better, two things happen: more people show up, and the products get more ambitious. It’s one of the most reliable patterns in tech, and it’s worth spelling out, because it explains a lot about the AI moment we’re living through.

Let’s use video games as the example, because the story is clean.

Making a game in the 70s and 80s was brutally hard (I presume, I wasn’t alive back then). Tools were proprietary, knowledge was scarce, and only a select few talented people could ship anything. When the talent pool for a sector is small, the innovation coming out of that sector is small too. Innovation scales on talent. The more people who can build, the more complex problems get tackled, and products rely on innovation to be good.

Then game engines went open source and general purpose. Unity, Unreal, Godot. Tutorials for everything. Suddenly a teenager in their bedroom had access to the same foundational tooling as a professional studio. Millions of people started making games as a hobby, and a meaningful slice of those hobbyists fell in love with the craft and went pro. The talent pool didn’t just grow, it exploded. And when more talented people are available, studios can take on harder problems and more ambitious projects. That’s how we got Elden Ring, Breath of the Wild, Red Dead Redemption 2, Minecraft. If the industry had kept gatekeeping knowledge and tooling, we’d have spent a lot more years playing Tetris and Pong. Great games, but not very complex ones.

Knowledge is a compounding force. Someone builds something, it sparks an idea in someone else, who builds the next thing, and so on forever. And the bar never stops rising because we’re insatiable. When a company ships a game with mechanics nobody has seen before, those mechanics become the new industry standard, and going back is nearly impossible. Consumer expectations only move in one direction.

So what about AI?

AI doesn’t break this pattern. It continues it.

Because that’s all AI is: another tool. This is the cornerstone concept of technology. Every time a tool of the trade gets better, the products built with it get more complex and interesting. Animation software let studios make films that hand-drawn cel work could never have supported. Video editing software turned a job that required a cutting room and physical film into something that unlocks entirely new kinds of storytelling. IDEs, compilers, debuggers, version control: each one made the underlying work faster and less painful, and each time the freed-up space got filled with harder, more ambitious work. AI is the latest entry in that lineage, not an exception to it.

And it’s an incredible tool. It lowers the bar and removes gatekeeping. It raises the average complexity of what gets built: the scope of what a single person can pull off on their own is dramatically bigger than it was two years ago. That shift cascades upward. Solo builders ship things that used to need a team. Small teams punch at the weight of companies. Companies take on projects they would have shelved as too expensive. Work that took days takes hours now, and that’s a force multiplier we can use to reach new heights.

No CEO in the history of capitalism has ever said “I’m happy with our throughput, we’re doing enough.” Never. If a CEO could press a button and spawn 100 more workers with zero downsides, they’d press it every single time, because there’s always more to build, more to fix, more territory to conquer. Productivity gains don’t shrink ambition. They feed it.

And if AI were really the death of all current products, where are the clones? If anyone can vibe code anything, why isn’t the world flooded with Slack, YouTube, and Android competitors? Why aren’t we drowning in AI-generated games that people actually play at scale?

Because code was never the bottleneck for good products. Someone who couldn’t code before can now build a cute demo, and that’s genuinely awesome. But building a full, fun, intuitive product takes attention, resilience, grit, and real passion for the thing you’re making. The person who quit because “I can’t code” will find a new excuse a few meters down the road. AI didn’t remove the obstacle. It just moved it.

The real skill issue

You’ve probably heard the camera analogy (Or something similar): AI slop is like handing someone a professional camera and getting back blurry, badly framed photos. I think the analogy is wrong. AI isn’t one camera. It’s a whole bag of cameras and accessories, and the skill is knowing when to use the fancy new gear and when to reach for the trusty old film camera.

That makes the actual problems clarity and restraint:

  • Clarity: knowing when to vibe code and when to do it traditionally.
  • Restraint: being strong-willed enough to not pull the slot machine lever on something that benefits from your care, attention, and time.

To be clear, I’m not against generating slop. Not all code is critical, and anyone who tells you otherwise should reflect on why they became an engineer. Testing harnesses, benchmarks, throwaway dashboards, vulnerability scans: these are fantastic at 80% completion, and grinding out the last 20% adds almost nothing. Why bother?

The skill issue is mental, not technical. It’s the discipline to look at a task and say “this one deserves the harder path, because the result will be better,” and the judgment to know which tasks earn that call. The harder path doesn’t mean swearing off AI, either. Brainstorming with it, refining ideas with it, handing it the tedious contained tasks: all of that is immensely valuable, because it frees you up for the critical path.

In fact, you could argue that making elegant products is harder in the age of AI, precisely because the AI can do everything. Restraint is on you now, and restraint is not something humans are famous for.

Where this goes

Because AI lowers the barrier so dramatically, because being an expert is no longer a prerequisite for building something and finding out, I hope it leads us to more novel, more technically complex products on average than we have today.

The tools will keep improving. The bar will keep rising. And the need for curious, creative, tenacious people to meet it will rise right along with it. It always has.