Sheesh! To me, a lot of Photoshop know-how was about using layers and masks and learning to manually, cleverly extract parts of an image, edit them, and put them back together.

Qwen-Image-Layered: Layered Decomposition for Inherent Editablity

qwen.ai/blog?id=qwen... (edited)

This is a pretty interesting OS tool from Anthropic:
www.anthropic.com/research/pet...

Great data visualization from Our World in Data. Question what the media is peddling:

ourworldindata.org/does-the-new...

Don’t blame Trump, Musk and the other fascists, blame the morons who voted for them!

@cantwell.senate.gov

Where exactly are you today - at which protest can we find you?

@repkimschrier.bsky.social

Where exactly are you today - at which protest can we find you?

@jayapal.house.gov

Where are you today - at which protest event can we find you?

@murray.senate.gov

Where exactly are you today - at which protest can we find you?

@jayapal.house.gov @murray.senate.gov @cantwell.senate.gov @repkimschrier.bsky.social

We voted for you. Now we’re not hearing from you - what are you going to do about this sh*t show?!

It’s time for all of you to go - you failed us.

This is the kinda stuff I am looking for and want the media to get focused on: Trump voters getting screwed by recent government gutting. Listen to him whine and talk about contacting his representative to complain! Trump is his representative!
www.youtube.com/watch?v=N23K...

While still a proof of concept, the results suggest this could be a powerful new direction for language models - combining internal reasoning with traditional token generation.