This Week in AI: What Actually Matters for Designers
Designers™
AI Download · September 25, 2026
This Week in AI: What Actually Matters for Designers
Okay, so I did not think I'd be leading with a Google Docs feature this week, but here we are. Four things crossed my desk that are actually worth your five minutes, and none of them are the usual "we made the chatbot slightly smarter" news. I don't keep up with all of it either, honestly, so consider this the filtered version.
Gemini can finally make usable video, and it's already inside Google Vids
Google rolled out Gemini Omni 1.1 Flash inside Google Vids this week, bumping generated and extended video clips up to full 1080p with actual duration control instead of a fixed clip length. If you've ever tried to stitch together a before/after reel or a quick project walkthrough and gotten stuck with a three-second clip that cuts off mid-sentence, this is aimed directly at that problem.
Why it matters for you: this is one of the first times AI video generation has felt close to something you'd actually post, not just something you'd show your team as a novelty. I haven't tested it myself yet, so take that as cautious optimism, not a review. The takeaway either way is that AI video for real estate and design content is moving fast, and it's worth a look before your competitors get there first.
Gemini Notebooks now grounds itself in real sources, which is the honest way to do this
Google also expanded Notebooks in Gemini, and the bigger deal is a new feature that lets you ground AI writing in Google Docs on sources you actually picked, plus access to over 100,000 published books through partner libraries. Instead of the AI just generating something that sounds plausible, you can point it at real material and have it work from that.
This is the part I actually care about. It lies about what it can do sometimes, we've all seen it happen, so anything that forces the tool to cite something real instead of confidently making it up is a step in the right direction. If you're using AI to help draft client proposals, spec sheets, or anything with facts in it, grounding it in sources you trust is not optional, it's just good practice.
Midjourney's editor got a real organization overhaul
The September 16 alpha update to Midjourney's v8.2 editor added a grid view with hover previews for your source images, replaced the old Upscale button with a simpler Enhance option, and let you drop images straight into folders from inside the editor, even ones that didn't exist yet. Small stuff, but if you've ever lost a reference image in a sea of generations, you know why this matters.
For moodboard building and image curation specifically, this is a genuine quality of life upgrade, not a flashy new model. That's such a time saver when you're three hours deep into client concepts and can't remember which generation had the right sofa.
Adobe Firefly can now remove backgrounds from video clips, not just photos
Firefly added the ability to automatically identify the subject in a video clip and strip the background throughout, the same one-click removal you already get on still images, now working on motion footage inside the Firefly video editor.
If you shoot product or install footage on a busy job site and want a clean background for a social clip or a client presentation without pulling out a green screen, this is worth trying. I'd still check the edges on anything with fine detail like plants or textiles before you call it done.
The Moral of Today's Story
You don't need all four of these. Pick the one that solves a bottleneck you actually have right now, whether that's video content, cleaner research, moodboard chaos, or messy background footage, and start there. Everything else can wait.
Sources
A quick note on how this gets made: I use AI to help research and draft this post every week, then I review and edit it myself before it goes out. Being upfront about that is part of what I teach, so here it is.
