|
Today's Supernew Debrief: Google's Gemini 2.5 Flash Image is here, and the image generation capabilities are huge!
AUGUST 31, 2025
|
|
|
News You Can Use
Google Gemini 2.5 Flash Image is here, and their image generation is beating everything.
Gemini (basically Google's version of ChatGPT) has some decent general updates in version 2.5, but the best ones are around image generation, especially with character consistency, targeting transformations and speed.
What does that mean? You can now have Gemini swap or restyle clothing, hair styles, texture, products or backgrounds in a photo and it will keep everything else perfectly intact, like people's faces, objects or vehicles.
Most AI image generation tools up until now would attempt to recreate the entire image and would inadvertently change the look, but Gemini is able to essentially target just the part you want to change, and nothing else.
This was always doable in Photoshop, though it required some serious skill and quite a bit of time. Now, it can be done in seconds by anyone.
Beyond that, you can restore old photos, combine various images into one naturally, restyle an image, and even have it show you the view from a point on a map (pulling from Google Earth/Maps data). Check out some of the examples we tried this week:
|
Prompt: "Swap the lighting fixture in the second image to the lighting fixture in the first image. Keep other details the same."
Prompt: "Put this Bronco in the world of Tron."
Note: I've tried a very similar prompt to this with every AI model in the last year, and nothing has gotten anywhere near this good. Most models just default to a standard stock Bronco, and miss all the unique modifications of this one. Gemini nailed almost every detail while restyling the environment.
Prompt: "Change the red part of the controller (2nd image) to this glass texture (1st image). Keep all other details the same."
Prompt: "Combine the dog and man in these photos into a single portrait, where both are together outside."
Prompt: "Make these people look like they're at the beach."
Prompt: "Restore this image and bring it to life with accurate color."
Prompt: "What does the black arrow see? Make me a picture of its viewpoint."
That's all the news for today!
If you'd like to read Google's announcement, you can do that here.
If you'd like to try it out yourself, just open Google Gemini's web app at gemini.google.com.
This new image model is all anyone in the AI community is talking about this week, as we're all discovering new ways to use it. It's equally exciting and terrifying. With that, we have one more thing worth mentioning.
|
One last note:
As this kind of technology becomes more accessible, we'd like to remind you to be careful with what information you share online and with AI tools. You've just seen how a photo of you or someone you know could now be used to create something you don't want.
The possibility for someone to create deepfakes of your likeness and use it to scam a family member or worse is now much more doable.
That doesn't mean we should live in fear or never use these tools, but we should be aware of what we share online, and with these AI services. We should also make our families aware of the possible nefarious ways this could be used, and remind them not to trust everything they see. We need to verify carefully and use these tools wisely. Our goal is to help you do that, so if you have a family or friend who would benefit, feel free to forward them this email.
|
|
|
|
Tool spotlight
|
|
Google Gemini 2.5 Flash Image (aka Nano-Banana)
If you follow much in the AI community, you've probably seen the term "Nano-Banana" floating around a lot recently. That's just Google's codename for their latest image generation model; it gained so much popularity that it stuck. But the official name is Gemini 2.5 Flash Image. Anyway, you should try it!
|
|
|
|
Quote of the day
“
Seemingly Conscious AI (SCAI) is the illusion that an AI is a conscious entity. It's not - but replicates markers of consciousness so convincingly it seems indistinguishable from you + I claiming we're conscious. It can already be built with today's tech. And it's dangerous.
Why it matters: to be clear, there's zero evidence of AI consciousness today. But if people just perceive it as conscious, they will believe that perception as reality. Even if the consciousness itself is not real, the social impacts certainly are.
— Mustafa Suleyman, CEO of Microsoft AI
|
|
|
|
what we're working on
|
|
The ultimate AI training library by humans for humans
We've built an expansive multi-media resource library to help you make the most of AI in daily life and work. Check it out
|
|
|
|
ABOUT THIS NEWSLETTER
You're receiving this email because you subscribed to or were forwarded the Supernew Debrief. Supernew was started by Ryan Smith and Ben Peacock to be your go-to AI resource in life and work. Each week, we share the top AI news, tips, tools and resources we've discovered or created to help you stay ahead of the AI curve. We hope you'll join us.
FOLLOW US
SUBSCRIPTION PREFERENCES
Subscribe, update your subscription preferences or unsubscribe from the Supernew Debrief.
113 Cherry St #92768, Seattle, WA 98104-2205
|
|
|
