Simple AI apps are THE opportunity

THIS WEEK’S STORY

AI Innovation Moves Fast... But Here’s How You Stay Ahead

This past week, the pace of AI development has been insane (again).

Deepseek R1 is kicking ass, Pika 2.1 can now drop any character into AI-generated videos with stunning quality, Alibaba rolled out their new LLM Qwen 2.0 Max, and Perplexity’s Sonar API just unlocked web search. And guess what? This will be the rhythm every other week in 2025—constant breakthroughs.

But here's the thing: the real opportunity isn’t in chasing these models—it’s in building “GPT wrappers” around them (shoutout to @gregisenberg for making this so clear).

What used to be seen as “low status”, it is now THE status.

Consumers don’t care which model powers your product. They care about easy-to-use, intuitive, and familiar interfaces. The trick is designing beautiful, engaging experiences that keep users coming back for more.

💡 Tech advantage is temporary – interface lock-in is forever.

Let the AI giants duke it out to see who has the “best model.” Your edge is leveraging these models to create standout interfaces and loyal users.

The Untapped Goldmine: AI-Powered iOS Apps

Here’s the kicker: while everyone’s building web apps and desktop software, hardly anyone is focusing on iOS-native AI apps—apps with dead-simple, yet powerful AI functionalities and jaw-dropping UI/UX.

Why? Because developing iOS apps is a beast if you don’t have a tech background.

But that changes when you harness AI to build it for you.

Knowing how to create iOS apps, even without coding experience, gives you a huge edge in this market.

⏰ This is why I’m launching a community to teach you how to build your dream app from scratch—with zero coding required.

AI will handle the development. You just bring your creativity and ideas.

I’m laying out the structure and modules for the course now, and I’d love your input.

Fill out this quick form to tell me what you'd like to see inside. Let's build something amazing together!

APP IDEAS

Here are some ideas to get your creative juices flowing. You can always go to Perplexity or ChatGPT and ask to refine them or come up with new ideas:

THIS WEEK’S TIP

So how should you go about building your GPT wrapper in an iOS app?

You have two options:

1. Use the Swift community driven package for OpenAI public API, where you have everything pre-configured to build your AI interactions. The only problem is that it’s a “client-side implementation”, meaning that all the API calls are sent directly from the app, hence exposing your API keys + scalability issues. However, the package is easy to implement and perfect to launch/test fast.

2. Implement Firebase Cloud Functions to handle the AI integration from the server, meaning it is much more secure and scalable (this is my preferred choice)

“Okay but how do I do that if I don’t know nothing about it?”. Well, you detail the user flow and interaction and give it to CursorAI. For a calories tracking app analyzing the calories from a food picture, it could be something like this (use ChatGPT for this):

• User uploads a picture of their meal [Frontend]

• User taps on “analyze” [Frontend]

• HTTPS cloud function is triggered [Backend]

• Function must accepts image data (or URL) and a prompt [Backend]

• The prompt should specify that the user needs calories analysis [Backend]

• Call the OpenAI Vision API through the cloud function [Backend]

• Receive the response and format the output in JSON [Backend]

• Display the output in a specific format on the app [Frontend]

• Secure the function – set up rate limits and error handling [Backend]

After you input this into CursorAI, ask it to expand the plan and give you a specific implementation strategy. This will make sure that everything is taken care of. Remember to take it slowly and step-by-step, otherwise the AI will start to hallucinate.

The takeaway here is that for enabling AI functionalities in your iOS app, you need a Firebase Cloud Function (something that I didn’t know prior to building it!). Let AI handle the code.

For detailed guides and implementation info, I am building a dedicated community with learning material. Fill out this form to help shape the content of the course.

QUICK HITS

• DeepSeek-R1's release triggers $1tn tech stock crash as the Chinese AI model matches OpenAI's performance at 1/10th the cost, challenging U.S. dominance in AI development.

• Google launches Vertex AI RAG Engine, simplifying retrieval-augmented generation for developers building context-aware LLM applications, with integrations for Vector Search and third-party tools

• Vellum’s 2025 AI Development Report reveals 25% of enterprises now have AI apps in production, with customer service chatbots (51%) and document analysis tools (59%) leading adoption

• Tencent’s Hunyuan3D 2.0 generates detailed 3D models from single images in seconds, offering open-source tools for gaming and e-commerce developers

How did you find this week’s issue? Let me know your thoughts or questions. Just reply to this email – I read every message and use your input to make these emails better for you. Happy to connect with you on Twitter/X too

See you next week!

Filippo