The first seven blog posts on a Daily Prompt are all about getting started with different AIs.
- Get started with Copilot
- Get started with ChatGPT
- Get started with Gemini (or Nano Banana)
- Get started with Grok
- Get started with Claude (chat)
- Get started with Mistral (Le Chat)
- Get started with Meta AI
So, which one is best?
We obviously cannot declare a winner based on the very very very very simple prompt we used in our getting started posts and we don’t even think it makes sense to talk about which one is best.
First of all, the prompt we created left a million things up for interpretation or imagination and we got very different results from the various AIs. Secondly some AIs are focusing on image generations, other are focusing on code generation or research – judging Claude based on its ability to create an image is like judging a Shark on its ability to climb a tree.
Very much like humans have different areas of expertise, AIs have the same. On A Daily Prompt, we will test prompts and techniques on various models regularly to learn which models are best at what kind of assignments. Obviously things can change, but that will always be the case.
Even though we cannot declare a winner, we did find things that we didn’t like and things that we liked or even loved…
What we didn’t like?
Meta AI is confusing. The functionality is different depending on where you use it and it is unclear whether this is because our tests are running from Denmark or whether Meta needs to get their act together.
Microsofts long list of Copilots and we are actually not sure what we need to buy in order to have the Copilot we need. Licensing has never been a strength of Microsoft and Copilot doesn’t change that.
Many scammers creating apps or websites, which attempts to lure users into using their website and tricking people into signing up for things they didn’t intend to. This is not something the various AIs can do much about, it is just bad behavior. We do not like that.
What we liked?
Claude doesn’t do images, but instead it creates the code to create the scene you are looking for. This also points to the focus of Anthropic – they are in it for the pros.
Microsoft is building AI into workflows in many products and they are also very focused on the agentification of their products. Agents presents a dramatic increase in productivity for the entire company, not only individuals.
ChatGPT has an OK interface and with the newly released mechanism to branch a chat, it gives the user the ability for ChatGPT to forget some of its detours.
What we loved?
The way Grok presented not just one option, but an infinite list of options almost – we found that pretty cool.
The way Meta AI presented the images and allowed for modifications like 3D, retro, cookie dough etc. – we found that very cool as well.
Geminis availability on my Android phone with a click of a button
Mistral represents Europe in the AI game and we actually think they are doing a very good job. Of all “magical” images, Mistral LeChat was one we liked the most.
So what now?
Based on the series of very very different images generated, we think it is fair to say that if we had a specific image in mind, we should have expressed ourself more clearly. If we had given the same task to 7 individuals, they would also have given me 7 very different images as well, and some people might even also have written code to draw the image:-)
That is what this blog is all about, how do you write prompts, which gives you what you expect (if you have expectations). We will explore many different use-cases for AI and discuss how to write the right prompt and what level of trust you can have in the answer from AI.
Note, that all AIs can make mistakes, but so can all humans. If you use the right prompts/questions and set the right context, you have a higher chance of getting useful answers.
This post was written on September 12th, 2025 – things might/will change…
Until next time
A Daily Prompt
A Daily Prompt 
Leave a comment