OpenAI has lobotomized DALL-E to the point that it is virtually worthless for 
continuity.  I doubt they've done much better for switching between scenes of 
motion pictures.  You can't do such simple things as storyboards or graphic 
novels.  The way they achieved this feat of engineering?

By closing off the only way of getting a face to be the same recognizable 
character between frames -- which is to specify it as a well-sampled celebrity. 
They yammer at you how this is some sort of violation.  So what do you do to 
get it to render a face consistently between frames?  You describe the 
character in detail and hope that it will do something reasonable.  Here, for 
example, is what I was limited to doing for a storyboard I wanted to do:

https://twitter.com/jabowery/status/1755378830426915008 
<https://twitter.com/jabowery/status/1755378830426915008/photo/1>

https://twitter.com/jabowery/status/1755382127510564904


------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T4ad5d8c386d0e116-Mc34b9d32018972272eb7c796
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to