Now that the fog has cleared from this year’s #inktober I can finally assess my Starryai experiments. Safe to say this year was an utter disappointment. It seems the AI focused more on “style” templates to base artwork when given fewer descriptor terms. In the past it at least created something unique vs something generic when “style templates” weren’t an option.

Sure I could use more descriptor terms & build prompts but I wanted to see what an AI could interpret with fewer words with the word of the day. This year it seems it can understand/comprehend far less than before.