No doubt, we’re miles ahead from the first mainstream image generating model, DALL-E 2 (Apr 2022). When DALL-E 2 hit the market, I remember it was awful with letters and human hands, in particular. Image/Video generating models are getting better and better every month, but there is still a ways to go. Especially to make it easy for the “everyday” user.
I think a fun app/service will be to take any movie or commercial and replace the characters with yourself and your friends. Here’s my attempt to do this for a short clip from the Matrix, after Neo completes his initial combat training and says: “I know Kung-fu”
Hilariously, even though I explicitly included in my prompt “Don’t change the audio”, KlingAI changed it and I spit out some random language 😝😂🤣🤣 (Chinese?).
I’ll keep following the evolution of image/video. For the everyday user, I’ve definitely had fun turning images into videos. I’m sure professional image/video folks have benefitted, maybe creating cheap B-roles or motion graphics. But it’s not quite mainstream mainstream yet, if you know what I mean.
Engaged in a Perpetual Search for Knowledge