AI startups — Differentiation, defensibility, and moats.

Can AI startups building on top of the same models differentiate and build strong moats?

We’re about to see an explosion of AI startups doing “the same thing”. Many already been launched, with many more to come.

This is because AI is now accessible to developers in the form of open source AI models and APIs. Developers build services making AI technology available for SMBs, startups, and enterprises.

So if all these services are built on top of the same AI models and APIS, how are they differentiated? And can they build strong moats?

Twitter has different opinions about it.

AI is now accessible.

I wrote an article that building on top of the latest AI models is one of the biggest opportunities for startup founders.

A.I. models are the new App Store

The TLDR of the article is that the AI models that a few years ago would cost millions to develop, are now available in the form of APIs and open sourced models.

And there’s a big difference between the two.

As these APIs and opens source models are available to developers and startups, the competition will be big. Many startups will be launched with the same offering, adding a layer of UI/UX on top of the AI models.

So if many startups are going to be launched offering the same things, how can they differentiate? And most importantly, how can they build defensibility?

Strong companies require strong moats, and those building on top of the AI models will need to develop them as well.

Model based vs API based

Differentiation at the application layer — is it enough?

The application layer means much more than the UI/UX. It also includes the development of the brand over time, including positioning in the industry with good marketing strategies.

Many startups will differentiate with their marketing strategy, using different acquisition channels. It also includes the differentiation in price.

The marketing and pricing strategies compound over time, building the brand reputation. And reputation is a strong moat.

The UI/UX is one of the key differentiators why startups will win or lose customers. But I want to clarify what I mean when I say this:

UI — How it looks.

UX — How it works.

UI/UX can be copied. What can’t be copied is the insights required to develop a great UI/UX, solving the customer painpoint in the easiest possible way. And that, is a key differentiator.

Building custom AI models

At CreatorKit, we are building our custom AI model: The CreatorKit Diffusion model.

This enables us to create AI product images that are actually better than any image generated with using the OpenAI API, Stable diffusion, or any other service so far.

We believe that the only way to provide better results on the generations is by developing our custom model trained with our data, and optimized with deep learning algorithms to our specific goals.

There are going to be hundreds of competitors launching, so we’ll see how it plays out!