In August 2023, we published an article titled “The Rise of AI Rappers and Future Challenges.” Almost a year later, these early warnings have arisen. It’s not the way everyone expected.
The work examined the flood of AI rappers, also known as API rappers. This is a tool that layeres a more user-friendly interface than models such as CHATGPT, Google Gemini, Claude and more. Most of them didn’t do anything groundbreaking under the hood. A powerful model has been made easier to use.
Then came the warning shot. In April 2024, Openai CEO Sam Altman notified the startup: If you’re just wrapping the GPT-4, “We’ll Steamroll You.”
He wasn’t bluffing. At the time, several startups were spinning up wrappers that looked like clever prompts dressed on the product.
But fast forward, it’s starting to look like both Altman – and we may have missed something.
The ground is changing.
AI Foundation model makers are losing ground
The rise of Chinese AI startup DeepSeek and the rise in price wars among basic model providers such as Openai, Gemini and humanity have revealed a shift in value.
As infrastructure costs drop and models become more difficult to differentiate, investors are beginning to realize that real opportunities lie in AI applications, not in the models themselves. This is a similar shift to how companies like Salesforce and Stripe build large businesses by focusing on software tiered on top of cloud infrastructure.
Open source models are also gaining traction, accelerating this shift in focus from the underlying technology to useful apps.
Think Salesforce or Stripe. The infrastructure allowed its existence, but it was the application layer that won.
That’s where AI rappers come.
The rise of AI apps
With regard to the large-scale language models and all the attention the companies that build them get, the true value of artificial intelligence is beginning to change. It’s no longer about who has the most advanced models. It’s about the people who are building the most useful applications on top of them.
Early headlines, centered around model makers like Openai and Google, are quietly moving towards app creators. This transforms raw AI into a tool that actually uses it. These builders now offer more tangible value and are attracting more attention.
As basic models become cheaper and more accessible, the edge that once belonged to model developers is fading. The focus is on the software layer that real-world use cases come into play.
Owning a model was previously the Holy Grail. However, with many high performance models currently on the market, the real advantage lies in the way you use them. And companies building practical, user-facing apps are beginning to take the lead.
Just as software has eaten the world, AI apps are beginning to eat AI itself. As the performance level of your model occurs, the most value is generated by apps that provide real utilities and a powerful user experience.
The real value of AI is shifting from the basic model to vertical AI apps
They were laughed at first. Products like Prplexity, Replit and Abridge have been ruled out as interfaces that are riding on other people’s technology. But while model providers are busy covering themselves to stay relevant, they still stand, but often thrive.
“The only way to compete in AI is to train these web-scale models in advance that can help you raise hundreds of millions of dollars and solve all the problems under the sun, and it was the only game in the AI town. “Very quickly, people actually realized that value moves up the stack.”
That view is becoming more common. Big names like Microsoft spent billions of dollars on model development. But the edges they gave are fast and narrower. Model performance is leveled, switching costs are low, and developers are increasingly moving towards open source options.
However, the app is one with a sticky user experience and clear use cases. They have a heart, heart, wallet. Andreessen Horowitz partner Bryan Kim put it like this:
“[Wrapper] Just kind of thing means it feels less thoughtful. It feels like you’re giving this little package around what’s built. In contrast to what it really means, “I’m going to understand the customer’s problem. I’m going to marry this and provide a solution to what you’re trying to achieve.”
That idea shapes how products are built into the valley. The code-centric wrapper, cursors are growing at a fierce speed. Build apps without the need for a CS degree. This is not about flashy technology. It’s about making something that people can actually use.
Calvin Chin of the E14 Fund is called this “atmosphere.” He imagines a world where there is just an atmosphere where people trust tools to do their jobs without having to inspect every detail.
“I love the phrase vibe coding, because in fact I think it points to me… in this new way we interact with these systems, we don’t necessarily interrogate everything they’re ongoing,” Chin said. “As the models improve and the products built on these products improve, we gain the activity of other kinds of atmospheres in the economy. So it’s an atmosphere, atmosphere, atmosphere, atmosphere, atmosphere, and we’ll trust the model more and more.”
So, what about Openai?
The GPT-4O is still ahead of its accuracy, nuance and overall experience. However, the lead is thinner.
Cost issues: The Deepseek R1 and others are rolling out models that are about as good, but much cheaper. If the GPT-4o really deserves a premium, companies are rethinking it.
Switching is easy. There is no need to rebuild the entire stack to switch providers. A few lines of code, and you’re good.
Open source is catching up. Projects like MixTral offer almost never before impressive results.
Meanwhile, the rapper is quietly building an empire.
AI rappers are quietly building an empire
Concept AI, confusion, and PoE prove that interfaces (how people interact with AI) are sticky. They focus on simplifying the experience rather than reinventing the wheel. And it’s working. While Openai, Google, and humanity are pouring billions into their infrastructure, these apps capture user loyalty by solving real problems.
At the end of the day, the model is just an engine. What users care about is the car.
And what makes great vehicles? They are the ones pulling ahead.
🚀Want to introduce the story?
Submit your stories to TechStartUps.com in front of thousands of founders, investors, PE companies, tech executives, decision makers and tech leaders.
Please attract attention
Source link
