Here are the 4 tips that got me started with Generative AI.
- Start with image generators – you quickly get a motivating impression of how prompt engineering works
- Program a «GPT» with Open AI: It can do a lot, you «program» with normal German or English.
- «Program» with no-code tools: Select predefined modules and combine them to create entire program functions including interfaces to external programs. By simply dragging lines.
- Program «right» with the help of Generative AI. It’s easier than you think!
1. Start with image generators - they quickly produce motivating results
Use Midjourney or other image generators such as DALL-E or Ideogram to create images or illustrations.
It is relatively easy to achieve appealing results very quickly, see the cover image of this article.
Look at sample images from other users in Midjourney or other image generators.
The text entries (prompts) that led to the results are usually listed below the images.
If you can produce appealing images, you have developed a good understanding of prompt engineering: You change or sharpen your text input based on the AI’s results.
Over time, you will find that while Midjourney’s images are amazing, you lack control over the creative process. Then you will want to look at ComfyAI and visual language models like Stable Diffusion or Flux. But don’t start there, it’s a bit complicated.
2. Use your English skills to program with GPT – without programming knowledge
Open AI not only offers its chatbot ChatGPT. You can also create your own «programs» that work like a personal chatbot.
Open AI calls them GPTs.
Here I have published a guide on how easy it is to program with GPTs:
- You describe in simple language what you need
- GPT asks for the desired name and icon of the program
- GPT then programs a corresponding application in the background
My description of GPTs is practically more complicated than programming your own GPT…
Programming a GPT makes it possible, among other things, to upload files for analysis and receive corresponding reports or even assessments.
This can be taken even further with so-called no-code tools.
3. "Programming" without knowing how to code with no-code tools
No-code tools such as Make.com or Zapier enable various functions to be linked together.
For example, I automatically translate my travel reports from Reisememo.ch via an interface to DeepL for my English travel blog Travelmemo.com
This is achieved by simply selecting predefined functions such as interface accesses (API).
The desired functions are «programmed» by dragging a connection.
That’s it…
No-code tools are very powerful. However, if you want to solve more complex tasks, actual programming is the way to go to build your data pipeline.
Thanks to Generative AI support, this too has become very easy.
4. Programming with the support of Generative AI
Generative AI such as ChatGPT from Open AI is based on their instance of a Large Language Model LLM.
Because programming languages are also «languages», many LLMs are not only proficient in creating text, but also in programming software code.
Accordingly, you can ask ChatGPT for programming code that executes certain functions. In the context of manipulating text and data, the programming language will usually be Python.
However, if you need a script for the 3D program Unity, you will get it as C# and for embedding code in Microsoft products it will be VisualBasic.
None of this should intimidate you, because you simply copy/paste the generated code into your program or run it on a web platform such as Google Colab or Databricks (Microsoft) or Anaconda (open source).
Here I explain how to get your own code running, using Anaconda as an example.
Conclusion for programming novices
Achieving inspiring results has now become very easy, even without programming knowledge.
For example, I read photos from my Flickr photo gallery, have ChatGPT comment on them in the style of an influencer and publish them on Instagram.
Everything is fully automated using Python scripts that ChatGPT has put together for me according to my needs.
So how do you prepare for the opportunities and risks of the coming workplace revolution?
