Yeah, last time a stealth model caught my attention like this it was o3 I believe
GENERATIVE AI
-
Model’s Realistic Aesthetic Compared to Glossy Alternatives
By
–
Fun model, especially for a realistic/candid look. I like how it seems capable of producing messy-looking, slightly imperfect shots, unlike something like Flux where everything is so glossy and pristine. https://
x.com/higgsfield_ai/
/higgsfield_ai/status/1940096729212465357
… -
Strange Cities: Style Curation in Generative Image Models
By
–
More strange cities.
— Ethan Mollick (@emollick) 1 juillet 2025
(Roughly similar prompts, but the ability to radically alter style makes curation and selection much more important) pic.twitter.com/1mTGMX2MkUMore strange cities. (Roughly similar prompts, but the ability to radically alter style makes curation and selection much more important)
-
Best lip-sync mode for video generation using adhere_1
By
–
If you care about lip-sync, I'm finding that `adhere_1` is the best mode to use. Also, you might need to trim a few frames off the start of the video as there can be a slight morphing effect. https://t.co/uLxeWkhdiT pic.twitter.com/ukNIDhaAVr
— fofr (@fofrAI) 1 juillet 2025If you care about lip-sync, I'm finding that `adhere_1` is the best mode to use. Also, you might need to trim a few frames off the start of the video as there can be a slight morphing effect.
-
AI tools for video restyling and modification via Flux and Luma
By
–
Some new things
— fofr (@fofrAI) 1 juillet 2025
1. Easily restyle the first (or last) frame of a video via kontext:https://t.co/tVAxC3zRJV
2. Apply the style to the whole video via luma:https://t.co/tudFiP7Drq pic.twitter.com/6fFBuv7FNlSome new things 1. Easily restyle the first (or last) frame of a video via kontext: https://
replicate.com/flux-kontext-a
pps/restyle-video-frame
… 2. Apply the style to the whole video via luma: https://
replicate.com/luma/modify-vi
deo
… -
Early look at future of AI Studio platform
By
–
An early look at the future of AI Studio (home page)
-
Midjourney Video Creates Weird Artistic Styles Exploration
By
–
Strange cities.
— Ethan Mollick (@emollick) 1 juillet 2025
(I find working with Midjourney video to be really interesting, the ability to develop weird styles especially) pic.twitter.com/Gmhhxy1y3qStrange cities. (I find working with Midjourney video to be really interesting, the ability to develop weird styles especially)
-
Praise for Model’s Realistic Skin and Training Quality
By
–
love this model. the skin, the realism, it's really solid. props to the team behind it. training something like this is not easy
-
Copyright and Transformative Use in AI-Generated Content
By
–
It's important to distinguish non-expressive use (inputs not even covered by Copyright, which some AI novices claimed) and non-expressive purpose (output that "Adds No New Creative Content", e.g. Books) that can show 'transformative use' after prima facie infringement was found.
-
Legal Analysis: AI Training, Copyright Infringement, and Fair Use
By
–
No, Brian is either senile or trolling. Sony was about contributory infringement, Fair Use analysis was thus done for private viewers. The Google Books case was deemed "non-expressive purpose" but their reproductions obviously contained expressions hence prima facie infringing.