All posts by jnack

The scarily beautiful animation of Sincitium

AI/ML, Illustration

Side note: “Macrófago” is 100% the best word I’ve learned all week.

AI filmmaking turns a (creepy, fun) corner

AI/ML

This is the first time I can recall watching a genuine narrative (not a handful of gee-whiz demo shots) made with AI & not really caring about the production details. We’re turning the inevitable corner where it’s just the quality of ideas & narrative that’ll matter—not so much how the proverbial sausage was made.

See yourself from a new angle in Google Photos

3D, AI/ML, Google Photos, Photography

Get some fresh perspective from our amazing teammates in research:

Today we are announcing a new approach to fix scene alignment after a photo was taken. Our method, now available as part of the Auto frame feature in Google Photos, uses machine learning (ML) models to understand the scene and its spatial layout and uses generative AI to imagine the photo from that new perspective. In contrast to classical photo editing, our method interprets a photo as a 3D scene — think of a real moment frozen in time — and change the camera position automatically within that space.

How to get the most from Nano Banana

AI/ML, Nano Banana

My new teammates have posted a series of detailed tips & tech specs (e.g. you can upload as many as 14 images together with a prompt). Check it out!

1. Introduction to Nano Banana

  • The Models: Overview of Nano Banana 2 (powered by real-time web search) and Nano Banana Pro (built for high-end reasoning).
  • Core Strengths: Deep reasoning capabilities, accurate visual rendering, and premium features like text rendering and upscaling (2K/4K).

2. Technical Specs at a Glance

  • Context Windows: Up to 131,072 input tokens for Nano Banana 2.
  • Versatility: Supports multiple aspect ratios (from 1:1 to 21:9) and up to 14 reference images in a single prompt.
  • Safety: Built-in SynthID watermarking and C2PA credentials for responsible AI use.

3. Best Practices for Prompting

  • Be Specific: Focus on concrete details regarding subject, lighting, and composition.
  • Positive Framing: Describe what should be there (e.g., “empty street”) rather than what shouldn’t.
  • Director’s Perspective: Use cinematic terms like “low angle,” “bokeh,” or “aerial view.”

4. Five Powerful Prompting Frameworks

  • Image Generation: Using the [Subject] + [Action] + [Context] + [Style] formula.
  • Image Editing: Utilizing “Semantic Masking” to change specific parts of an image via text.
  • Real-Time Data: Leveraging web search to create visuals based on current events or weather.
  • Text Rendering: How to get legible, localized text in over 10 languages within your images.
  • Creative Direction: Advanced tips for controlling lighting (e.g., Chiaroscuro), camera hardware (e.g., GoPro vs. Fujifilm), and film stock.

5. The Creative Ecosystem

  • How to combine Nano Banana with other models like Gemini (for prompt engineering), Veo (for video keyframes), and Lyria (for AI soundtracks).

Photoshop, 3D, and redemption

3D, AI/ML

“Being early is the same as being wrong.” — Marc Andreessen, Vol. ~900

We put 3D into Photoshop nearly 20 years ago, and it got used by nearly 20 people total, lol. For many of the past several years, it was on the team’s “gotta throw overboard, as soon as we can find time” list—but happily that time was never found.

I am so glad to see this foundation now finding a meaningful niche, and I have high hopes for its generative future. Posing a person or thing directly is so much more intuitive than trying to precisely describe an outcome via prompt, and simple 3D manipulation + generative rendering could well deliver game-changing best of both worlds.

Canva’s new Magic Layers converter is really impressive

AI/ML, Design

As generative imaging models like Nano Banana get increasingly adept at rendering text-heavy layouts, the ability to convert those layouts into native text/image compositions is of course hugely valuable for editing. Check out Canva’s new Magic Layers feature:

I couldn’t resist trying it out with a silly infographic I made using the new ChatGPT image model, and dang if it didn’t do a pretty a great job:

“LooseRoPE” promises super intuitive illustration & compositing

AI/ML, Illustration

Man, it must be nearly 20 years ago that we started envisioning drag-and-drop-simple composition and compositing in Photoshop—back when gradient-domain painting & blending was the emerging hotness. After plenty of false starts, could these simple interaction patterns finally become mainstream? Maybe! I must know more of this witchcraft:

A love letter to the Chicago lakefront

Design

My mom and her sisters (all with the skin tone 255/255/255) spent way too many days turning themselves into rotisserie chickens here over the years. That’s one of millions of stories, ranging from shipwrecks to chicanery, plucky birds to an African American rodeo scene, that have unfolded along Chicago’s amazing lakefront. Having grown up in Illinois & visited hundreds of times over the years, I really enjoyed this tour from WTTW & Geoffrey Baer:

Recent 3D hotness from Apple & Microsoft

3D, NeRFs & Splats

Apple’s LiTo can generate Gaussian spalts with realistic view-dependent rendering:

Meanwhile Microsoft’s open-source Trellis tech promises super fast 2D-to-3D conversion:

You can now verify Google AI-generated videos in the Gemini app

AI/ML

You can now check if a video was edited or created with Google AI directly in the Gemini app.

Just upload a video and ask something like, “Was this generated using Google AI?” Gemini will scan for the imperceptible SynthID watermark across both the audio and visual tracks and use its own reasoning to return a response that gives you context. For example, it might say: “SynthID detected within the audio between 10-20 secs. No SynthID detected in the visuals.”

Uploaded files can be up to 100 MB and 90 seconds long.

Scout with Maps, animate with Veo

AI/ML

Check out this super cool mashup between Google Maps & my new product, Veo (video generation):

The team writes,

With Maps Imagery Grounding, a film studio can use a laptop to quickly visualize a scene at a specific place, like Washington Square Park in New York City—before scouts ever set foot on set. It’s easy to use: just type a prompt like “generate an image of a futuristic spaceship hovering in front of the Washington Square Arch” into the Gemini Enterprise Agent Platform and enable grounding with Google Maps Imagery in settings. In seconds, you can storyboard your creative vision with an accurate image—and you can even use Veo to animate the scene.

Google Photos adds touch-up features

Google Photos

My first week at Google in 2014, I was asked to send a company-wide dogfood email about automatic teeth whitening in Photos. We got draaaaaged hard by offended Googlers, & the feature never launched.

The important lesson: cosmetic stuff can be welcome, but only as opt-in, not as AI-imposed judgement of your appearance. I used to liken the difference to my wife walking through a department store & being offered a makeover by a beauty products salesperson—versus that person simply grabbing her & forcibly applying the makeup (!).

In any case, now you can opt into using these features in Photos (only on Android at the moment, it appears):

You could have a steam train…

Illustration

Despite—or perhaps because of—growing up without MTV (I know, the Gen X horror...), I’ve always had a real fascination with the video for Peter Gabriel’s Sledgehammer. Check out its rad zoetrope picture disc incarnation:

And, because why not, it’s Friday & you deserve nice things, here’s the original vid:

“AI will never suffer from bipolar disorder and autism like me”

AI/ML

Spending four minutes listening to Diplo’s thoughts on how art will be made going forward, and specifically on the value of quirky, messy, world-experiencing humans will be a good use of your time, I promise. The machine needs us ghosts.

“A rare look at how Hollywood is already using AI”

AI/ML

I’ve been sending this video to friends & family to explain what the heck it is I actually, y’know, do for a living. (It’s somehow related to enabling all this!)

Here’s a good summary from Gemini:

  • Digital Clones for A-listers (0:33–1:56): The Creative Artist Agency (CAA) is helping actors create and store secure digital doubles of their likeness and vocal inflections. This serves as a “vault” to protect their intellectual property and assert rights against unauthorized use.
  • Deep Voodoo’s AI Innovations (2:15–3:54): Founded by Trey Parker and Matt Stone of South Park, this studio uses proprietary facial scanning and AI to perform tasks like real-time de-aging for projects like the TV series Before and Billy Joel‘s recent music video.
  • Production Efficiency and Ethics (6:03–7:40): Director Darren Aronofsky and filmmaker Eliza McNitt utilized Google’s Veo 3 model for the short film Ancestra. AI allowed them to create complex cosmic visuals and even recreate a newborn baby digitally to avoid the ethical concerns of filming with a real infant.
  • Commercially Safe AI Tools (8:00–9:10): Asteria Film Company, co-founded by Natasha Lyonne and Bin Moser, focuses on building “commercially safe” AI models trained strictly on licensed materials to avoid copyright infringement, emphasizing that learning to use AI is an essential skill for modern filmmakers.
  • The Human Element (4:48–5:13): Despite the rapid evolution of AI, industry unions like SAG-AFTRA emphasize that human performers bring a unique, special quality to projects that algorithms cannot replicate, advocating for guardrails to ensure AI serves as a tool for creators rather than a replacement.

Veo & the making of “Ancestra”

Miscellaneous

Honestly, in taking my new role at Google & working to bring Veo and other models to creators, it’ll likely be hard to focus on the more boring bits (which, as with every job, will certainly be there) when storytellers like Darren Aronofsky & Eliza McNitt are pushing the limits of the tech & all I want to do is dive in up to my eyeballs. 🙂 But, as they say, that’s a good problem to have, and I look forward to learning more over time.

Meanwhile, check out this look into the making of ANCESTRA, made by Eliza (and team) about her own birth:

Here’s the film itself:

No Star Wars? No Photoshop.

Design

The last time I visited Industrial Light & Magic, Russell Brown & I grabbed lunch with Photoshop co-creator John Knoll. As they’d just retired a bunch of bulky rendering hardware, John was busily removing the fascia (adorned with Imperial logos) and adding decorative blinkenlights, creating some pretty exceptional décor for his office.

I was reminded of this seeing Russell share this 1-minute history of how John’s work at ILM proved to be crucial in his & Thomas’s creation of Photoshop:

Here’s the full episode

00:00 Cold Open — AI, Creativity & The Big Question
00:50 Welcome to Creative Outsiders
01:09 Introducing Russell Brown (“Doc”)
02:00 Photoshop Origins: ILM, Star Wars & The Abyss
06:55 The “Holy Sh*t Moment” — Taking Control of Images
11:10 From Rub-Down Type to Digital Creativity
12:05 Where Creativity Comes From
16:00 Becoming the Best at What You Love
18:15 Enter AI — Tool or Threat?
24:00 The Future of Photography & AI Workflows
30:15 Creating Films with AI & Storyboarding
34:00 The Ethics of AI in Photography
37:00 AI for Pre-Visualization (Not Replacement)
43:00 From Photoshop Fear to AI Fear
44:30 Why Russell Shoots on iPhone
48:00 Simplicity, Constraints & Creativity

And just in case you’re curious, here’s John recreating the first demo of Photoshop, some 20 years after the fact (which is itself now 16 years ago, OMG…):

Big news: I’m back at Google!

Idle Philosophizing

Hey gang—I am beyond delighted to say that I’m returning to Google, taking a Cloud AI PM role focusing on generative media!

As Paul Simon told us, “These are the days of miracles and wonder”—and I wonder at my amazingly good fortune getting to help shape these miracles.

Ever since 2000, I’ve focused my PM career on “unblocking the light,” helping people make the world more beautiful and fun. From Photoshop to Google Photos to M365, I’ve loved learning what truly matters to creators. Nothing beats zeroing in on real needs, then marshaling some big giant brains to deliver everything from big breakthroughs to crafty little mint-on-the-pillow delights.

Returning from last fall’s Adobe MAX, I summarized attendees’ vibe as “Overwhelmed, But Optimistic.” Now the pressure—and privilege—is to turn that optimism into action.

There’s so much I don’t yet know about this role—but what I know for sure is that I can’t do it alone.

I know I need you.

As we all navigate this bewitching, bewildering time, let’s stick together. Please keep me honest, grounded in knowing just what you need—and what you don’t. That way I can advocate accordingly, helping Google focus on exactly what’ll benefit you most.

Questions & ideas for collaboration are always most welcome: [last name @ employer dot com]. And in the meantime I’ll keep sharing my most interesting finds on the ol’ blog—especially in the burgeoning AI/ML category.

And with that, friends, here we go!

Photo history: How NASA trained moonwalkers to shoot

Photography

I love behind-the-scenes little insights like these. Click or tap as needed to see the full post:

Phota launches, promising maximum identity preservation

AI/ML, Nano Banana, Photography

Phota—about which I expressed some initial misgivings, given its ability to rewrite memories—has launched Phota Studio & their API. From what I can tell, it builds upon a Nano Banana foundation and adds personalization that relies on uploading dozens of images of each individual in order to maximize identity preservation:

With Phota, for the first time, you can generate, edit, and enhance photos while keeping your identity intact, every time.

We’re not building a generic foundation model. We build personal models about you, and about the people and pets around you. At the center are profiles, built from your personal album that learn the details of your appearance that make you recognizable as yourself: how you smile, your eye color, and how your face looks from different angles. Your personal model is private and only used by you.

Here’s a quick thread in which I tried inserting myself into a couple of images, using both Phota’s model (which depended on my uploading 30+ images of myself) and just Nano Banana straight out of the Gemini app:

Inside the design of “Project Hail Mary”

Design

I love love love the attention to detail that Phil Lord and Christopher Miller brought to the film. Check out the lengths they & their crew went to on everything from devising rotating lights for the inter-ship tunnel (conveying constant rotation) to nailing film grain. And I love the exuberance & generosity of creators in sharing so many insights into design & process.

FreePik enables 3D photo shoots

3D, AI/ML

I love seeing progress like this: upload a product pic, convert it to 3D, and photograph it on a virtual set:

Runway debuts Multi-Shot

AI/ML

Here’s a fun, ultra-simple way to turn an image (or just a prompt) into a short, multi-shot narrative:

Just for fun I fed it this image…

…and this prompt (based on an all-too-true story):

A family of Lego people and their dog gaze around Yosemite’s most iconic vista, then reminisce about that time they got stuck there in the snow in their VW van, expressing hope that they don’t get stuck again!

Check out the results:

I Love(art) to move it, move it…

AI/ML, User Interface

I’ve long quoted James Ratliff, the super sharp designer behind Adobe’s Project Graph (who’s recently decamped to Figma), in nicely phrasing how the process of generating & refining ideas generally starts broad/declarative (searching, prompting) and moves towards fine-grained methods (selecting, moving, etc.):

I see an increasing number of tool & model creators mixing modalities—even in the Gemini Super Bowl ad featuring a mom & daughter drawing a simple circle to show where they’d like to add a dog bed.

I’m eager to check out Lovart’s take on the possibilities, especially for animation:

Update: Here’s a look at the UI, in which you can move & scale the selection rectangle, as well as the before & after images:

Spline enables agentic 3D creation

3D, AI/ML

“3D scenes, websites, games, apps,” promises Spline. “Describe anything and Omma builds it for you in seconds.”

Omma combines code generation (LLMs), 3D AI mesh generation, and Image generation all in one place for you to build and ship. Deploy to production, assign custom domains, and more.

Sora is the new Peach

AI/ML

Ten years ago (!), the embryonic social app Peach suddenly blew up on the scene—only to molder shortly thereafter. Adam Lisagor tartly predicted that outcome right after Peach debuted:

I’m reminded of this upon hearing that OpenAI has bailed out on Sora, which they launched just a few months ago. In a way I’m not surprised—check out how interest in the tech spiked & then rapidly cratered—except that just a couple of months ago Disney signed a billion-dollar deal to use it. ¯\_(ツ)_/¯

Luma UNI-1 promises layered creation

AI/ML

When can we get this (or equivalent) into Photoshop??

On a conceptually (though not necessarily technically) related note, the LICA dataset may help model makers train layered generation:

Photoshop drops Rotate Object!

3D, AI/ML

Speaking of spinning right ’round, check this out:

Check out another view, from Paul Trani:

Runway promises 100ms (!!) HD video generation

AI/ML

Five years ago, I spent an afternoon with a buddy watching Disco Diffusion resolve a weird, blurry, but ultimately delightful scene over the course of 15 minutes. Now Runway & NVIDIA are previewing generation that’s a mere ~90,000x faster than that. Ludicrous speed, go!!

Pass the Craic pipe: It’s St. Paddy’s!

Photography

Now throw your shoulders back and go effin’ nuts. 😀

And for some more blog-appropriate content: Here are some fun pics & vids my son Henry & I captured on Saturday during SF’s wonderfully diverse & quirky St. Patrick’s Day parade:

Bonus: here’s a gallery of Irish wolfhounds, if you’re into that kind of thing. I couldn’t quite get these good boys to align like Cerberus, so I resorted to telling Gemini my hopes & dreams—as one does.

Tips: Getting great text from Nano Banana

AI/ML, Nano Banana, Typography

Structuring your prompt well turns out to be key in avoiding garbled text. As the presenter says, “It’s not about writing more. It’s about writing in the right order.” Check out this brief overview.

In this tutorial, you’ll see how to use Nano Banana Pro and Kling 3.0 Omni together to solve one of the most common pain points in AI product video: text that blurs, warps, or drifts mid-motion. We’ll walk through a practical workflow for maintaining legibility and visual consistency in product shots, so your labels, logos, and copy stay clean from the first frame to the last.

Antigravity 360 drone spins me right ’round

Photography

Hey, remember the pandemic? We sure made some impulse buys then, didn’t we?

For me it was Insta360’s bizarre, modular 360º camera plus the elaborate mounting kit that promised to strap its shards onto the top & bottom of my DJI Mavic, enabling some magical, drone-less captures. Suffice it to say the thing was a complete POS—dysfunctional even as a handheld action cam, much less as a bunch of theoretically interconnected pieces thousands of feet in the air.

And yet… who doesn’t love the promise of capturing immersive footage that enables crazy post-processing camera moves? Insta’s on it, releasing their first 360º drone, the Antigravity A1:

Some cool details:

With Antigravity’s proprietary FreeMotion technology, the drone — together with the Vision goggles and Grip controller — enables an immersive flying experience that feels both natural and intuitive. Pilots can fly in one direction while looking in another. This level of immersion enables more freedom to explore. The 360 immersion doesn’t end just because the drone lands — recorded footage can be viewed in 360 over and over again, letting users discover new angles every time they watch.

Using AI to save pets

AI/ML, Illustration

Long dog walks are for nothing if not visualizing whatever silliness pops into my head—which today happened to be our puppy Ziggy becoming an impossible object called a “Ziggule.”

I shared this with my cousin Alicia, who does a tremendous amount of work sheltering & rescuing dogs in Austin, and she requested a portrait of their current foster pooch (Tesseract). I was of course all too happy to oblige:

As it happens, folks at Google have had the same idea, and they’ve been putting Nano Banana to work helping zhuzh up pics of shelter pets in hopes of helping them find their forever homes. Let’s hear it for using AI & old-fashioned human creativity for good!