🐙 We just published the 8th version of #PeerTube !

New video player, collaborative channel management, improved import system…
With each update, PeerTube improves to better meet your needs!

In a new blog post, we review these new features, as well as the past year and our plans for the future.

Check it out on Framablog!

👉 https://framablog.org/2025/12/09/peertube-v8-manage-your-videos-with-your-team/

@peertube the image look really AI-generated 🫣

EDIT: IT IS NOT

@blueluma It's amazing how the big industry has brainwashed the audience. I now collect this type of accusation no longer monthly, but weekly. And as a digital painter with more than 25 years' experience, I understand why. In fact, my work (and other, billion) has been used to train those AI models, so AI paint like us. But please stop accusing work that you are not 100% sure about. When real humans are behind it and have spent full days creating that imagery, it hurts a lot.

@peertube

@davidrevoy Interesting to see a bit of the process, and I think one of the issues might be the compression-induced "vaseline filter" on the image in the original post? Your post has a much less compressed version that's much sharper and thus more obviously not AI (My initial verdict was "probably human, low confidence" due to the compression making it hard to confidently determine "human" until your post)

(But I can see where the original accusation was coming from, even if I would never have made it myself: with the heavy compression in the original post it's much harder to determine if possible "tells" are actually tells, or just Compression...)

@becomethewaifu You've got a good point there, Emelia. It looks like the illustration in the original post was edited, with lots of reduction or enlargement filters or JPG 'smoothing' compression. A lot of the texture has been smoothed out, and the AI really goes to town on smoothing things (these 'AI gen' image techniques are Denoiser on steroids, after all). The full resolution ( https://www.peppercarrot.com/0_sources/0ther/framasoft/hi-res/2022-11-25_Peertube-v5_by-David-Revoy.jpg ) reveals really the details and brush strokes.

@davidrevoy In this case it's more that the smoothing hides the other tells than being a tell itself, if that makes sense? Like in the original I can see the crisp shape of mini-Sepia on the lower screen, but in the compressed version it just looks like a detail-less smear (which is exactly what "AI" tends to do for those types of details) but it isn't clearly a nonsense-blob, so I couldn't confidently say it was a tell, and not just the compression smearing it.

Also, at least on my main monitor (but not my secondary; will check my actually color-calibrated laptop later), the color grading on the light in the upper left happens to match the "AI piss filter" color pretty closely. The screen having a very 'pure' blue makes it clear that it's not the piss filter, since that's always over the entire image when present, but some people might not realize that.

EDIT: got left and right backwards again...