As a Data Scientist and an enthusiast of AI technologies, I recently delved into an intriguing video by Matt Wolfe titled “What AI Image Generator Should YOU Be Using?” In this video, Matt provides a comprehensive comparison of various AI tools based on different attributes.
What AI Image Generator Should YOU Be Using❓

Here is the image I performed OCR on
Intrigued by his analysis, I decided to apply Optical Character Recognition (OCR) to extract the data from the video and then organized it into an Additive Tree. For those unfamiliar, Additive Trees are an excellent tool for visualizing similarities and differences between various elements. They use a correlation matrix to illustrate how closely related different items are: the closer they are in the tree, the more similarities they share; conversely, the farther apart, the less they have in common.
While it’s common to use Additive Trees for analyzing data from hundreds of responses, even with a single source like Matt’s video, valuable insights can be gleaned.
For example, my analysis highlighted a strong correlation between the cost of these AI tools and their capabilities for text in images and price, as indicated by the minimal distance between these two factors in the tree.
However, there’s a noticeable disparity between text integration and the quality of illustration output, as evidenced by the considerable distance between these two aspects on the tree.
Similarities of Attributes / Functionality
On a side note, the color coding in the tree is purely for visual aid and does not signify any specific data pattern. It’s just a way to help discern the various connections and distinctions more clearly.
Simmilarities of AI Image Tools
By transposing the data, the same thing can be done to show the relationshps between the AI tools.
Below you can see that FireFly2 is different than all the others (because is is on it’s own branch) but Ideogram and Google and Dalle3-Bing and Dalle-3-chatgpt are very simarl because they are all clustered on their own branch.
Again, this is based solely on Matt’s data but it does help give you an idea of the similarity between attributes and tools.