What Makers Told Us: Danioros on the two hardest moments in DO NOT WAKE HER

Meem: Seedance 2.5 on Higgsfield AI

Meem's 30-second found-footage clip, made with Seedance 2.5 on Higgsfield AI, could easily pass for a real early-2000s home video. The only giveaways are a few odd prompted actions, like moving a peg on an empty washing line.

Meem's 30-second clip is a captivating found-footage home video with all the flavor of the early 2000s, and it is today's Video of the Day on thegenerated.com. One continuous shot tracks a young woman as she puts up her washing, waters the plants and walks downstairs into the street.

A couple of moments give it away, but only in the prompted actions. For some reason she rearranges a peg on a line with no washing on it, empties what looks like a dog bowl with no apparent purpose, and when she reaches the street a bike sounds a car horn before she steps out.

Despite those unusual moments, you could easily mistake the whole thing for a real home video. Even when her face is hidden by the sheet, or as she walks away from the camera, she remains consistent. A really nice take on the style.

Meem says it was made with Seedance 2.5 on Higgsfield AI. If you have made something in this spirit, tag us so we can take a look.

Every clip is credited to its maker. If one is yours and you want it removed, message us and it comes down.

Source: Meem (mehvishs25) on X, 14 Oct 2026.

YouTube Music starts labelling songs made 'primarily' with generative AI

YouTube Music now labels songs made mostly with AI, using partner declarations and its own detection where nobody declares. YouTube has not said how accurate that detection is or how a wrong label gets fixed.

YouTube Music started labeling songs "created primarily using generative AI" on Oct 14. YouTube's example is a track where a model made both the composition and the vocals. The tag sits on the player page, and tapping it opens a "How this was made" panel that reads "Made with AI: This audio was altered or fully generated using AI." YouTube says the labels will also come to the main YouTube app.

The detail that matters is who decides. YouTube says it is working with music partners and "using our own detection tools" to find AI tracks. Its help page lets partners declare each track as fully AI, partly AI or no AI, and warns that with no declaration YouTube "will rely solely on those signals and may designate content as fully or partly Gen AI."

That same help page lists its own examples, and one of them is ours: a song with 100% human audio whose music video visuals were made with a text to video model counts as partly AI. Pitch correction and automated AI mastering count as no AI.

What stays open is how accurate the detection is, how a wrong label gets corrected, and whether partly AI tracks show a visible label at all. YouTube says it is up to partners to interpret the definitions.

In practice, if your short or music video runs on an AI score or AI vocals, declare it yourself when you deliver, so the label you get is one you chose.

Source: YouTube blog, YouTube Help, Oct 14, 2026.

What Makers Told Us: @wtry1102, Rough 3D box mock-ups for positioning and camera angles helped the AI generate natural fig

A filmmaker explains how rough 3D mock-ups helped an AI model generate natural fighting movement in an action scene.

WTR (@wtry1102) told us that using simple 3D box mock-ups for positioning and camera angles helped a multi-character action scene generate real fighting movement instead of stiff results. They shared this after we asked how the method held up on a more complex scene.

The lesson for other makers is that you don't need a complex 3D setup to guide an AI-generated shot. Even rough shapes can help with character positioning, camera angles, and movement.

Their words and work are credited to them. If this is yours and you want it removed, message us and it comes down.

Source: WTR (@wtry1102) replying to us on X, 13 Oct 2026.

Jordan Daniel Chesney: I'm heading to LA to speak at the CapCut World LA event , October 24th!

Jordan Daniel Chesney announced his CapCut World LA talk with a fictional comic action short about his own arrival. The post does not name the tools behind it, and it doubles as a promo for the event.

Jordan Daniel Chesney is heading to LA to speak at CapCut World LA on 24 October, and he announced it with a really fun, realistic fictional feature about his own arrival, brought to life with comic book action and comedy.

From a giant CapCut logo blocking the highway to Jordan sliding face first down the side of the event venue before he walks in, even the unbelievable moments are made believable in this short. Jordan himself keeps a consistent and realistic likeness throughout.

It even opens with a "just one more gen" line, which puts it firmly inside the generative universe from the first frames. Thoroughly enjoyable.

Jordan says his session will cover the new dawn of AI filmmaking and as many tips and tricks as he can fit in. If you are going, or you have made something in this spirit, tag us so we can take a look.

Every clip is credited to its maker. If one is yours and you want it removed, message us and it comes down.

Source: Jordan Daniel Chesney (jordandchesney) on X, 14 Oct 2026.