Good morning. ByteDance dropped Seedream 5.0 Pro and it can slice a finished image into 10+ editable layers, which is kind of insane for anyone building thumbnails or title cards. Google Photos added an AI Video Remix tool that restyles your clips in a couple taps with zero editing skills. And a "drone tribute" to Ronaldo that had the whole timeline crying turned out to be fully AI-generated. Which of these are you actually adding to your workflow this week? Hit reply and tell me. We cover all three below.
🎨 Seedream 5.0 Pro Can Slice Your Image Into 10+ Editable Layers
ByteDance launched Seedream 5.0 Pro on July 8, the top tier of its Seedream image family sitting above 5.0 and 5.0 Lite. It's a multimodal image model built for actual production work, not just pretty one-off generations. The headline upgrades: layer separation, precision editing, dense infographic layouts, and native text rendering in 14 languages.
The one that matters most for creators is layer separation. Feed it a finished image and it breaks the thing into 10-plus independent layers (background, subject, text, effects) and hands them back as transparent PNGs. It even inpaints the areas that were hidden behind your subject, so you can swap a parrot for a peacock without the background falling apart. Pair that with point, box, and anchor selection editing, where you change one element while everything else stays locked, and you get one-click tweaks instead of regenerating the whole frame.
For anyone producing thumbnails, title cards, or reference frames at volume, editable layers change the math. You stop treating each generation as final and start treating it as a working file. It also keeps the 5.0 family's built-in web search and deep reasoning, so trend-aware prompts pull live context instead of guessing. It's live now in Dreamina with free daily credits, plus API access if you want to wire it into a pipeline.
In Partnership
10x the context. Half the time.
Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.
Google Photos Will Now AI-Remix Your Videos in a Couple Taps
Google added a Video Remix tool to Google Photos, tucked inside the Create tab next to collages and the Photo to Video feature. It's powered by Gemini Omni, Google's new multimodal model, and it turns raw phone clips into stylized versions without any manual editing. Pick a video, pick a template, hit Generate, done.
The templates cover cinematic relighting to rescue a dark clip, background swaps, and artistic looks like watercolor, raw sketchbook, and oil painting. It can generate sound too. This is template-driven, so you're not color grading or masking anything yourself, you're picking a vibe and letting Gemini Omni handle the transform.
Be clear on what this is though: it's built for the average person fixing a dim birthday video, not serious production. Results get glitchy with fast motion, busy backgrounds, or shaky footage, and there's a daily generation cap unless you're on a paid Google AI plan (it's rolling out to Plus, Pro, and Ultra subscribers first). The real story is distribution. When one-tap restyling lives inside the app where billions already store their footage, the bar for "good enough" quick edits just moved, and that's a quiet nudge at the CapCut crowd.
That Ronaldo "Drone Tribute" Everyone Cried Over Was Fully AI
After Portugal lost 2-1 to Spain and got knocked out of the 2026 World Cup, likely Ronaldo's last, a video went megaviral showing a drone show over Madeira, his home island. Drones formed the Portuguese flag, the words "Obrigado, 7," and a glowing figure doing the siu celebration over the bay. People shared it everywhere as a real farewell. It was not real.
Fact-checkers traced the clip back to an AI creator who posted it on TikTok and openly said it was AI-generated, even offering to share the prompts. No drone show ever happened, and there was zero local coverage of an event that, had it been real, would have been everywhere.
This is the part worth sitting with. The craft was good enough that millions read a text-to-video clip as documentary footage of a real event. The tells were tiny: reflections that didn't line up, phone screens in-frame showing different footage. For creators, that's roughly the current ceiling of the tools you're already using, and it's a reminder that "wait, is this even real" is now the default question on every viral clip. Also yes, the siu but make it drones is a genuinely great concept. Someone go build the real one.