431–440 of 1001 posts
Was surprised to see it be able to render a decent comic illustrating an unemployed Pac-Man forced to find work as a glorified pie chart in a boardroom of ghosts.
Earlier quoted context omitted.
That makes sense[1] but it prompts the obvious question: does this style write it as typeö then? 1: Though personally I hate it, I just cannot not read those as completely different vowels (in particular ï → [i:] or the ee in need ; ë → [je:] or the first e here; and ö → [ø] or the e in her )
No. Firstly because it is spelled “typo.” Secondly you typically use the diaeresis to tell the reader to not confuse it with a similarly spelled sound or diphthong. So it tells a reader that “reëlect” is not pronounced REEL-ect, “coöperate” is not COOP-uh-ray-t, and “naïve” is not NAY-v.
Earlier quoted context omitted.
Why would you consider this a good prompt?
[flagged]
We have enough people complaining about Simon Willison's pelican test.
do they have anything similar to SynthID, or are they just pretending that problem doesn't exist? I know this is probably mega cherry-picked to look more impressive, but some of the images are terrifyingly realistic. They seem to have put a lot of effort into the lighting.
At least they aren't pretending that a solution exists.
One of the images in the blog ( https://images.ctfassets.net/kftzwdyauwt9/4d5dizAOajLfAXkGZ7... ) is a carbon copy of an image from an article posted Mar 27, 2026 with credits given to an individual: https://www.cornellsun.com/article/2026/03/cornell-accepts-5... Was this an oversight? Or did their new image generation model generate an image that was essentially a copy of an existing image?
This is hilarious. Seems like kind of a random image for a model to memorize, but it could be. There is definitely enough empirical validation that shows image models retain lots of original copies in their weights, despite how much AI boosters think otherwise. That said, it is often images that end up in the training set many times, and I would think it strange for this image to do that. Regardless, great find.
Earlier quoted context omitted.
So do you think there will be a better image model in a year?
I'm honestly unsure what could be improved at this point. Consistency? So it fails less often? Based on the released images, (especially the one "screenshot" of the Mac desktop) I feel like the best images from this model are so visually flawless that the only way to tell they're fake is by reasoning about the content of the image itself (ex. "Apple never made a red iPhone 15, so this image is probably fake" or "Cost…
That's because you're focusing a little bit too much on visual fidelity. It's still relatively trivial to create a moderately complex prompt and have it fail miserably.
Even SOTA models only scored a 12 out of 15 on my benchmarks, and that was without me deliberately trying to "flex" to break the model.
Here's one I just came up with:
A Mercator projection of earth where the land/oceans are inverted. (aka land = ocean, and oceans = land)The quality of the text is really impressive and I can’t seem to see any artefacts at all. The fake desktop is particularly good: Nano Banana would definitely slip up with at least a few bits of the background.
Are camera manufacturers working on signed images? That seems like the only way our trust in any digital media doesn't collapse entirely.
One of the images in the blog ( https://images.ctfassets.net/kftzwdyauwt9/4d5dizAOajLfAXkGZ7... ) is a carbon copy of an image from an article posted Mar 27, 2026 with credits given to an individual: https://www.cornellsun.com/article/2026/03/cornell-accepts-5... Was this an oversight? Or did their new image generation model generate an image that was essentially a copy of an existing image?
Given the recency of that image, it is unlikely it is in the training data and therefore I would go with oversight.
Earlier quoted context omitted.
> What else? I used to have an assistant make little index-card sized agendas for gettogethers when folks were in town or I was organising a holiday or offsite. They used to be physical; now it's a cute thing I can text around so everyone knows when they should be up by (and by when, if they've slept in, they can go back to bed). AI has been good at making these. They don't need to be works of art, just cute and sill…
You are kidding, right? It's good that my friends don't make a coffee date feel like a board meeting (with an agenda shared by post 14 working days ahead of the meeting, form for proxy voting attached).