Week 1 Discussion

What surprised me about how AI handled text or media generation?
I was pleasantly surprised by how well the AI would generate the same style of image if using the same prompt (with a slight change here or there) that specifies the style/aesthetic desired. I will say any image generation that included text, this was a weak spot where the text would be blurry, warped, incomprehensible, and not configurable to match theme fonts.

Text to audio generation was also really fast. This could save us hours if not days of work (and re-recording is a beast all on its own)! I love the variety of voice options that are available and most voices seemed very realistic and could add emphasis in the right places. 

What challenges did I face when trying to get AI to produce the results I wanted?
As mentioned above, any image that was AI-generated that included text came out warped - it was easier to ask it not to include any text (which was hit and miss with whether it listened or not), or to crop out any section that included text. Also the custom block is NOT included in any AI editing tools (yet?) so that limited my experimentation a bit.