KI Tools: Your Digital Toolbox for Everyday Use

A contribution from the KI Kantine by Michael Busch

“Which program should I use?” – I hear this question often when people are starting out in the KI world. The answer is: There isn’t one single program. Just like in a canteen, there are different stations – a salad bar, a main course, a dessert. You choose where it tastes best.

The Thinking Team: ChatGPT, Claude & Perplexity

If I want to understand something – a new framework, an interface, or even legal questions – I first turn to my “thinking team”:

ChatGPT is my constant companion and all-rounder. He knows practically everything and is the first port of call for most questions. A starting point.

Claude is sometimes more empathetic in its responses, but tends to get a bit off-topic. Perfect if you need a more considered perspective.

Perplexity is my tool for current information. It can search the web brilliantly, but you need to know how to ask it the right questions.

The Building Department: No-Code Tools

When I’m tinkering and experimenting, I turn to no-code tools:

V0 (Vercel) is like Figma, only the AI provides the code immediately. You essentially draw an app, and the AI creates the code. You can use the result immediately and even publish it.

Lovable is even more clever. I input a very structured prompt, and the AI builds the entire workflow with reusable components. It even asks if it should connect me to a database. I usually have Claude or ChatGPT create the prompt for this.

A practical example: I simply created a scr A screenshot of a website I liked was taken, and I told V0: “Build me something like this from the image.” The result was impressively good.

The Professional League: Programming Environments

For fine-tuning, I use proper programming environments with built-in AI:

VS Code was long my standard, but since Cursor AI and Windsurf came out, I’ve switched. These tools have made the AI much more user-friendly. They are essentially siblings of VS Code – they took the open-source code and improved it.

In these environments, you can really discuss with the AI how to best approach a problem. You first plan together, then say, “I think we’re agreed – please do it.” And the AI works it out.

The Iterative Workflow

Important to understand: I don’t turn these tools on one after the other, but rather they are all like tools on my digital desk.

• Planning: Thinking team
• Building: No-code tool
• Fine-tuning: Programming environments

But there are constantly recurring iterations needed. Sometimes I have to go back to the thinking team because something fundamentally doesn’t work. Then I pass the change on to the no-code tools.

The AI Kicker Tournament

Sometimes I even pit the tools against each other. I ask ChatGPT, “What do you think of the prompt Claude just gave me?” This is like an AI kicker tournament – I turn various knobs and see which result is best.

My Recommendation for Beginners

If you’re just starting out, I’d recommend starting with ChatGPT. You can download it as a desktop app or just use it in the browser.

Once you’ve had enough conversations with ChatGPT, definitely try a no-coder. It’s a lot of fun! You enter some “nonsense” and wait and see. Often it gets stuck and doesn’t progress – but sometimes something truly exciting emerges.

Conclusion

The AI tool landscape AFT is like a well-organized toolbox. Each tool has its place and its strengths. The trick is choosing the right tool for the right task – and knowing when to switch between them.

There isn’t one perfect AI tool. But there’s the perfect mix for your needs.

---

The AI Kantine is a project by Michael Busch, who regularly serves up fresh AI dishes – with a pinch of humor and a side of skepticism.