Tootfinder

Opt-in global Mastodon full text search. Join the index!

No exact results. Similar results found.
@Techmeme@techhub.social
2026-08-20 06:21:17

A look at Backstory, an experimental AI image authentication tool from Google DeepMind, offered for testing to journalists, researchers, and other fact checkers (Andrew Deck/Nieman Lab)
niemanlab.org/2026/08/googles-

@padraig@mastodon.ie
2026-09-20 13:45:09

I think @… has lost it lads.
(I know it's a supplement to the main headline about the overfishing of mackerel... just out of context, it is random)

Screenshot from The Journal

Category: "THE DAILY POLL"
Headline: "Do you like mackerel?"
1hr ago, popularity (??) 3.8k

There is an image of an open can of mackerel?
@toxi@mastodon.thi.ng
2026-07-18 16:56:00

Two small new prints from this afternoon, this time not Kallitype but contact printing on silver chloride FB paper, developed with Adox MCC. The increase in depth and detail is just wonderful and the feel of the final results an instant throwback to my early darkroom adventures ~40 years ago... 😍
The pictures are of two of my favorite places in recent years. Larger prints will be available for purchase in September.

Phone capture of two small B&W prints with black border, depicting beautiful alpine mountain landscapes.

The top image shows a path in a high alpine valley leading into the distance past a small lake, surrounded by steep rock faces. Moody dark clouds forming.

The bottom image shows a large boulder within a group of swiss pines (also growing on top) in a talus field. Further in the distance the glaciated peaks of the Bernina range in bright sunshine.
@lukem@hachyderm.io
2026-09-19 18:40:25

TIL in 2024, during some AI image contest, someone submitted a real non-generated photograph. And they won. Then they got disqualified.
artnews.com/art-news/news/phot

@Mediagazer@mstdn.social
2026-07-07 20:35:14

Meta launches Muse Image in Meta AI, Instagram, and WhatsApp, and previews Muse Video, the first media generation models from its Superintelligence Labs (Meta)
ai.meta.com/blog/introducing-m

@michabbb@social.vivaldi.net
2026-07-14 01:47:52

📊 Results: text context produced 50 successful games at ~9 minutes average. Image context produced only 23 successful games at ~20 minutes average. Damage numbers were comparable or slightly better with images, but the runtime gap was drastic.
💸 Token usage: 306 million tokens for the text runs vs 213 million for the image runs — with roughly half as many completed games.

@bilbo_le_hobbit@mamot.fr
2026-08-17 13:13:34

Dis Bilbo, c'était quoi ton moment préféré ?
Y en a un paquet, évidemment, mais il y a ce moment après le troisième contact quand la pression retombe, que les gens autour sont béats et joyeux. Et ce petit rituel du selfie solaire instauré il y a 20 ans dans le désert égyptien. Un moment de partage avec tous les voisins qui sont repartis avec un croissant de soleil en projection sur leur t-shirt ou dans la main. C'était chouette, vraiment...

Homme heureux et béat en chemisette et t-shirt, ce dernier étant une relique d'une précédente éclipse en Zambie en 2001. Sur le t-shirt, un croissant de soleil orangé.et l'image projetée vient d'une petite lunette astro sans filtre solaire dirigée vers le t-shirt. Attention ! Ne jamais mettre l'œil à l'oculaire d'un instrument pointé vers le soleil s'il n'est pas muni d'un filtre adapté !
 cascade réalisée par un astronome amateur, à ne pas réaliser chez soi sans connaissance préalable.

Trump’s Super PAC Starts to Unload Its War Chest,
With New Group Booking $47 Million
A recently created Republican group tied to the president’s super PAC
is pumping cash into half a dozen key midterm races,
with the largest amounts going toward Senate contests in Michigan and Ohio

@aardrian@toot.cafe
2026-07-09 15:40:54

I also saw that uncredited, context-free photo in my timeline.
This entire thread is a good explanation of the many ways that post was bad form:
mas.to/@meganL/116888106293249
Give credit, give context, don’t do inspo porn.

@michabbb@social.vivaldi.net
2026-07-14 01:47:52

🧠 The reasoning: text tokens are discrete, drawn from a fixed vocabulary of ~50,000 entries, each mapping to a fixed point in embedding space.
Image tokens are continuous vectors of a thousand floating point numbers, so they can pack far more information into the same space.
🎮 The test: a tower-defense game where the AI plays in JSON mode without graphics, run for hours with two context variants — the same game state as a plain text string, and as an image rendered via PixelPipe.…