MC.72: Wrapper's Delight

Hey everyone,

There's been growing criticism lately about AI applications being "just wrappers" around models like ChatGPT or Claude. This dismissive label has become increasingly common, suggesting these products offer little innovation beyond repackaging existing technology.

But is this perspective missing the forest for the trees?

What Makes a "Wrapper" Anyway?

When we label something as "just a wrapper," we're often overlooking a fundamental truth about software development: everything is built on something else. We all use APIs, libraries, and frameworks. If you dig deep enough, you can always claim that "X is just a wrapper for Y."

The more important question isn't how something is built, but what problem it solves.

Consider this: even a simple UI change to ChatGPT's web interface could dramatically improve results for users. If that's the case, does it matter that it's "just a wrapper"? What matters is the outcome.

The Evolving AI Ecosystem

We're witnessing an interesting trend in the AI space:

1. Frontier model providers (OpenAI, Anthropic, etc.) are moving up the stack toward applications, trying to get closer to end-users

2. Application developers are using these APIs to create new use cases and demonstrate their vision for AI interaction

This tension is creating a vibrant ecosystem where no one yet knows the best way to use AI or what the killer interfaces will be.

The Value Beyond the Wrapper

Differentiation in this space can take many forms:

• User experience innovations that make AI more accessible or powerful

• Marketing and branding that connects with specific audiences

• Rapid iteration that allows testing ideas quickly before building deeper technology

• Specialized use cases that solve specific problems better than general-purpose tools

As we saw with the Manus situation a few weeks ago, cloning AI applications has become trivially easy. But this only reinforces the need to focus on value creation beyond the underlying model.

Embracing the Wrapper Mindset

Perhaps we should view wrappers more positively - they enable rapid experimentation, allow creators to test ideas quickly, and prevent us from being locked into a single way of using AI.

Some of these "wrappers" will evolve into billion-dollar companies, while others will remain side projects or prototypes. And that's perfectly fine.

In a world where copying is increasingly trivial, the value shifts to capital efficiency, brand strength, acquisition strategy, and most importantly, creating unique user experiences that solve real problems in distinctive ways.

So the next time you hear someone dismiss a product as "just a wrapper," remember: it's not about how it's built, but what it enables users to do.

Cheers, Olivier

This newsletter was created with Alter—AI that enhances rather than replaces your best thinking.

Like this newsletter? Forward it to a friend and have them sign up here.

Until next Thursday 🎉