Yes, It Already Exists. Yes, Everything's a Remix. No, That Doesn't Matter.

Recently I shared a hypothetical startup idea to a friend of mine, who asked "Doesn't that already exist?"

I was about to turn the conversation into a boring presentation about what I have learned and think about the landscape of startups... but I didn't want to ruin the chill vibe. So I just said "Please" (and yes, I've been re-binge-watching HIMYM... :))

Since I can't let a thought go, here's what I would have answered.

That question, "doesn't this already exist?" actually reveals something deeper: not all startups innovate in the same way. In my head startups can be divided into "levels of innovation." Some tackle straightforward issues within a narrow niche, while others push the boundaries of science itself. Applying "levels of problem-solving" can help us better understand where a new venture sits, what risks it carries, and what kind of impact it might have.

Of course, this is just one way to slice it - and I don't claim to have any real expertise in founding startups. But as I discussed in my Learning as a Meta‑Skill article, grouping things is not only fun and satisfactory, but also helps us compartmentalize knowledge and improve learning. This framework is meant to be a thinking tool, not a rigid taxonomy - many successful companies blur the boundaries between levels. So here it goes:

Level 1: Same Niche, Same Problem, Better Execution

Existing niche, existing problem, even existing solution - but improved execution

This is where you take an existing problem - say people need to take notes. Evernote was one of the first to popularize the idea of a digital notebook that synced across devices and let you clip content from the web. Later, Bear stepped in with a cleaner, more minimal design aimed at writers who valued Markdown and distraction-free focus. Then came Notion, which expanded the concept into an all-in-one workspace by blending notes with tasks, wikis, and lightweight databases. None of these reinvented the problem - people still just needed a place to write things down - but each refined the experience, added polish, and packaged it in a way that felt fresher and more useful.

Docker fits perfectly here too. Cgroups and containerization technology already existed in Linux, but Docker gave it better packaging, cleaner APIs, and made it accessible to developers who weren't kernel experts. Similarly, Zoom's success came from brilliant execution at scale, taking existing video compression and streaming technology and engineering it to work reliably for millions of simultaneous users with an intuitive interface.

Solutions typically stand out thanks to their execution: ease of use, nicer packaging, better integrations, more user-friendly UX.

This evolutionary approach is the sweet spot for many successful startups because the market is proven and user behavior is established. The easy part? You don't need to educate users about the problem or convince them they need a solution. The hard part? It's brutally competitive since barriers to entry are relatively low, the niche is typically already occupied by established players, and differentiation often comes down to minor improvements that competitors can quickly copy.

Level 2: Same Niche, Adjacent Solutions

Existing niche, existing problem - borrowing proven solutions from adjacent fields

Level 2 companies look beyond their immediate competitors but stay within the same broad industry. Take Airbnb. They didn't invent hospitality or booking systems. Instead, they borrowed the two-sided marketplace model that worked brilliantly in other parts of the travel industry (think Expedia connecting travelers with airlines) and applied it to personal lodging. Same industry, different specialty.

Uber did something similar. They took the taxi dispatch system that already existed and combined it with ride-sharing concepts from other transportation sectors. The innovation wasn't inventing ride-hailing; it was applying existing logistics solutions from freight and delivery to personal transportation.

Kubernetes also fits here - Google took their internal container orchestration solutions (brilliant engineering at unprecedented scale) and packaged them as a product for the broader market. The core orchestration challenges had been solved internally; the innovation was making those solutions work for everyone else.

The easy part here is that you're still working within familiar territory - users understand the basic need, and you have adjacent success stories to learn from. The hard part is convincing an established industry to adopt a foreign approach, often facing regulatory hurdles and entrenched interests that benefit from the status quo.

Level 3: Cross-Industry Solution Transfer

Taking proven solutions from completely different industries

This is where things get interesting. Netflix's breakthrough wasn't better video technology. It was applying the subscription model from magazine publishing to entertainment. They saw how magazines solved customer retention and revenue predictability, then transplanted that business model into a completely different industry.

Stripe follows this pattern. They took the merchant account system that worked for traditional retail and applied it to online payments, making it dead simple for developers to accept payments without dealing with banking bureaucracy.

What's easy about Level 3? The solution is already proven to work somewhere else, so you have a roadmap. What's hard? You need to deeply understand two completely different industries and convince people in the target industry that a "foreign" approach makes sense. Plus, the translation between industries often requires significant technical adaptation.

Level 4: Scientific Breakthrough Application

Solutions require diving into underutilized scientific principles

These companies succeed by commercializing scientific discoveries that haven't found widespread application yet. DeepMind's protein folding breakthrough with AlphaFold exemplifies this. They took computational biology research that existed in academic papers and turned it into a tool that could revolutionize drug discovery.

ChatGPT sits somewhere between Level 3 and Level 4. The transformer architecture and attention mechanisms existed in research papers for years, but OpenAI's innovation involved both scaling proven science and solving entirely new problems around training stability, safety, and human alignment that hadn't been tackled before.

The easy part? If you succeed, you often have a substantial moat because few people understand the underlying science well enough to replicate it quickly. The hard part? You need deep technical expertise, significant capital to bridge the gap between lab and market, and patience to work through problems that don't have Stack Overflow answers.

Level 5: Beyond Current Science

First discover, then innovate

Level 5 innovations require advancing the scientific frontier itself. Think early quantum computing companies like IBM's quantum division or Google's quantum supremacy work. They're not just applying existing science. They're pushing the boundaries of what's scientifically possible while simultaneously trying to build commercial applications.

SpaceX's reusable rocket technology operates here too. The fundamental engineering challenges of landing and reusing orbital-class rockets had to be solved from first principles. Existing aerospace science provided a foundation, but the specific problems of propulsive landing and rapid reusability required new breakthroughs.

What's easy? Virtually nothing, except maybe fundraising if you can prove the science works. What's hard? Everything else. You're simultaneously doing R&D and trying to build a business, often with no clear timeline for when either will work.

The Real Point

The key insight? When someone asks "Doesn't that already exist?" the answer reveals which level you're playing at. Level 1 says "Yes, but we do it better." Level 5 says "No, because the science didn't exist until we created it."

But here's what I really wanted to tell my friend, and what I've been thinking about more introspectively: why do people really create startups?

Well, sometimes the world needs another note-taking app. Sometimes it needs quantum computers. Both are valid ways to scratch the entrepreneurial itch.

For me, it's about creativity - the pure joy of building something new, even if that something is "just" a better note-taking app. It's about tackling larger, interconnected problems because even a Level 1 startup forces you to think about user experience, business models, marketing, operations, and a dozen other disciplines simultaneously. And yes, it's about autonomy - the fundamental human need to create, to solve problems (big or small), to have the freedom to pursue your vision of how things could work better and just to be able to Build Your Own Arcane.