Sponsored by
Welcome to Today’s AIography!
The big news of the last seven days came in twos. Two Asian video labs. Two American post-production tools. Two pieces of news that change what a working creator should test next.
Let me walk you through them in the order they actually matter.
In today's AIography:
Kling 3.0 hits native 4K. What it actually changes.
ComfyUI raises $30 million at a $500 million valuation
DaVinci Resolve 21 ships 9 new AI tools and a Photo page
Imagen Video lands in Premiere and Resolve
Sora's app went dark. What filled the hole.
Essential Tools
Short Takes
Video of the Week
Final Thoughts
Read time: About 8 minutes
THE LATEST NEWS
TL;DR: On April 24, Kling AI rolled out version 3.0 on OpenArt. It's the first major AI video model to make native 4K video from a single text prompt. No upscaling. No post-process. The picture comes out at 3840×2160 the moment the model finishes thinking. OpenArt reported beta signups doubled in the first ten hours. Multi-shot sequences up to 15 seconds, sound that lines up with picture, consistent characters across cuts, and lighting you can guide.
Key Takeaways:
Native 4K, no upscale step. First major model to deliver 3840×2160 directly from text. The "render at 1080p, upscale" workflow becomes optional, not required.
15-second multi-shot sequences with cut consistency. Same character, same lighting, across multiple shots in one generation. Solves the AI filmmaking pain point that broke most narrative work in 2025.
Audio that lines up with picture. Native sound generation timed to motion. Not a separate sync pass.
Beta signups doubled in 10 hours on OpenArt. Demand signal is real.
The unanswered question is whether it survives a real cut. Marketing footage looks great. First honest colorist tests will land over the next 7-10 days.
My Take:
Native 4K matters most when you're delivering for a 65-inch screen. Broadcast, festival submissions, anything aimed at a real-screen audience. You used to render at 1080p and upscale, and the upscale was usually fine — but it was a step that cost time and money across a project.
For social-native work the math is different. Phone screens cap out at numbers most viewers can't see. A 1080p clip you cared for in the cut beats a 4K clip you didn't.
The honest practitioner question is whether the native 4K actually holds up when you pull it into Premiere or Resolve and grade it next to camera footage. I don't know yet. Neither does anyone else. We're three days in. Test it on something you'd actually deliver and tell me what you saw.
Try This Now:
Sign up for the Kling 3.0 beta on OpenArt and run one prompt at native 4K.
Pull the output into Premiere Pro or DaVinci Resolve. Grade it next to a piece of your own camera footage. Look at faces.
Generate a 15-second multi-shot sequence with one character. Cut to the second shot. Watch what survives the cut and what doesn't.
If the 4K holds up at delivery, plan one paid project around it this month. If it breaks, log what broke and where.
TL;DR: The same day Kling shipped native 4K, ComfyUI announced a $30 million Series B led by Craft Ventures, with Pace Capital, Chemistry, and TruArrow joining. Total funding: $48 million. Post-money valuation: $500 million. The platform has 4 million users, 60,000 community-built nodes, and 150,000 daily downloads. It runs in production at Netflix and Nike.
Key Takeaways:
$500M valuation on an open-source workflow tool. Series B led by Craft Ventures. Total funding to date: $48M.
4 million users, 60,000 nodes, 150,000 daily downloads. The numbers compound network effects.
Already in production at Netflix and Nike. The studios shipping work you watch run their AI pipelines through it.
Visual node editor for AI pipelines. Drag a model node, connect to a prompt node, connect to output. A generator becomes a controllable production tool.
Open-source moat is the node catalog, not the platform. Same dynamic that made WordPress impossible to dislodge in publishing.
My Take:
The funding round will fade in twelve months. The 60,000 nodes won't. Network effects in open source compound, and every node a creator builds becomes a node every other creator can pull off the shelf.
Here's what this means for filmmakers and content creators. The open-source side of the AI pipeline isn't the rough alternative to the paid tools anymore. It's the production layer the paid tools have to compete with. If you've been holding off on learning ComfyUI because it looked too technical, the case for learning it just got stronger. Institutional money brings better documentation, more tutorials, more workflows you can copy.
The case for staying on paid tools is also still real. If your week is full and the $20-a-month workflow you have works, a tool you understand beats a tool you don't. ComfyUI's win doesn't require you to switch. It just means the people switching are getting more wind at their backs.
Try This Now:
Install ComfyUI from github.com/comfyanonymous/ComfyUI. Free, open-source, runs locally.
Pull one of the top community workflows from ComfyUI Manager and run it on a reference image you already have.
If you've been on Midjourney + Runway, run the same generation through a ComfyUI workflow and compare control. The difference will be obvious in 20 minutes.
Bookmark the ComfyUI subreddit for daily node releases. The platform's strength is the catalog, and the catalog grows fast.
TL;DR: Blackmagic Design released DaVinci Resolve 21 as a public beta on April 13. Free download on the Blackmagic site. Newsshooter's NAB hands-on dropped April 21. Two headlines: a brand-new Photo page that brings node-based color grading to still photographs, and nine new AI tools — most notably AI Face Reshaper and Face Age Transformer.
Key Takeaways:
The Photo page is new. First time a video tool brings node-based color-grading software (curves, qualifiers, power windows, full node editor) to still photographs inside the same project.
AI Face Reshaper. Change eyes, nose, mouth, eyebrows, face shape on a moving subject. Auto-tracking. Slider control.
Face Age Transformer. Aging or de-aging a tracked subject without a separate VFX pass.
Smaller adds: AI Blemish Removal, IntelliSearch (find clips by what's in them), CineFocus (change focal point in post), tethered camera controls.
Free public beta today. GA later in 2026.
My Take:
Face Reshaper and Face Age Transformer are the most interesting because they used to live outside the editing software entirely. You sent the shot to a VFX house. You waited a week. You paid for it. Now it's a slider in the program you already cut in. The first time it saves a continuity problem on a film you actually shipped, the math will be obvious.
The ethics conversations around face manipulation will write themselves over the next month. Larger publications will cover that beat. The practitioner question is when a tool like this saves a shot and when it creates more problems than it solves.
Try This Now:
Download the DaVinci Resolve 21 public beta from blackmagicdesign.com. Free.
Pull a take you would have abandoned for a continuity issue (face turn was wrong, eyes were closed, expression broke). Test Face Reshaper on it.
Run Face Age Transformer on a reference shot of yourself or a willing subject. The control sliders teach the tool's range fast.
Open the new Photo page with a still you shot recently. See if your photo workflow can move into Resolve.
TL;DR: Imagen Video left beta on April 21 with native plug-ins for Adobe Premiere Pro and DaVinci Resolve. The pitch: automatic baseline color grades — white balance, skin tones, sensor-mismatch corrections, and LUT-aware AI Profiles. The company says these run roughly ten times faster than the manual versions. The CEO calls it "the co-pilot every editor deserves." Editors keep creative control on top.
Key Takeaways:
First major NAB 2026 product to actually ship. Most NAB announcements are still "coming soon." Imagen is available right now.
Native plug-ins for Premiere Pro AND DaVinci Resolve. Not a stand-alone app. Lives where you already cut.
Automatic baseline grade in seconds. White balance, skin tones, sensor-mismatch corrections. Roughly 10× faster than manual.
LUT-aware AI Profiles. Layers on top of your existing LUT pipeline rather than replacing it.
Editors keep creative control. The grade is a starting point, not a finished pass.
My Take:
The use case for filmmakers and content creators: any project where you have a lot of footage and no time to grade it shot by shot. Documentary editors will see the math first. Music video editors with day-of-shoot deliverables will see it second. Anyone working with multiple cameras and inconsistent lighting will see it third.
The thing to watch: AI baseline grades have been promised for years. The reason editors haven't switched is that the grades looked machine-generated. Too clean. Too consistent. Missing the mood the colorist would have brought. Imagen is the first tool in this category coming out of NAB. Whether it survives a real grade is the question worth answering.
Try This Now:
Install the Imagen Video plug-in for Premiere or Resolve from imagen-ai.com.
Pull a multi-camera doc or interview project you graded by hand. Run Imagen's baseline pass on the same timeline. Compare.
Test it on mismatched-camera footage (phone + DSLR + mirrorless in one cut). That's where the white-balance math should pay off most.
Don't ship the auto-grade as final. Use it as a baseline. Build your look on top.
CREATOR ECONOMY
Sora's App Goes Dark — Where the Migration Went
TL;DR: OpenAI's consumer app for Sora went officially dark on Sunday, April 26. The API stays alive until September 24. Shutdown was announced March 24, so this is the calendar date arriving, not new news. Yesterday was the practical day workflow-builders had to migrate. The migration scattered across the field rather than landing on one replacement.
Key Takeaways:
Consumer app dark April 26. API alive until September 24.
Runway Gen-4.5 picked up the cinematic-quality scene crowd.
Veo 3.1 (free for any Google account, 10 generations/month, 720p, 8 seconds, watermarked) picked up the short-clip crowd.
Kling 3.0 (the new 4K version) picked up the long-form 4K crowd.
Pika kept its niche on viral effects (Pikaffects, Pikaswaps).
Seedance 2.0 held its lead for multi-shot stories with native sound.
The economics: $15M/day in compute against $2.1M lifetime revenue per public reporting. The shutdown was the predictable end of an unsustainable plan.
My Take:
There's no single replacement. The hole Sora left got filled by the rest of the field doing what they were already doing better. Pick the tool that fits the job, not the tool that has the most marketing budget this month.
The bigger lesson is the burn rate. $15 million a day against $2.1 million in lifetime revenue. That's the kind of number that ends products no matter how good the marketing was. If you've been building workflows on a "freemium AI video" tool that hasn't shown a sustainable business yet, the calendar is shorter than you think. Diversify your toolkit.
Try This Now:
If you used Sora's consumer app, pick ONE replacement from the list above based on what you actually used Sora for. Don't try them all this week.
Migrate one in-flight project to the new tool and compare what breaks.
Document which generations you'd done that you'd want to keep. The Sora API window closes September 24 — pull anything you need before then.
Build your toolkit so no single product going dark blows up your week.
ESSENTIAL TOOLS
AI Filmmaking & Content Creation Tools Database
Check out the Alpha version of our AI Tools Database. We will be adding to it on a regular basis. Got a tip about a great new tool? Send it along to us at: [email protected]
Kling 3.0 (Beta). First major AI video model with native 4K from text. 15-second multi-shot sequences, integrated audio.
ComfyUI. Open-source visual node editor for AI pipelines. 4M users, 60K community nodes. Just raised $30M at $500M valuation.
DaVinci Resolve 21 (Public Beta). Free upgrade. New Photo page + 9 new AI tools including Face Reshaper and Face Age Transformer.
Imagen Video. Native plug-ins for Premiere Pro and DaVinci Resolve. Automatic baseline color grades, ~10× faster than manual. First NAB 2026 product to ship.
Veo 3.1 (Free). 10 generations/month for any Google account. 720p, 8 seconds, watermarked. Picked up the Sora short-clip crowd.
Runway Gen-4.5. Picked up the cinematic-quality scene crowd post-Sora migration.
SHORT TAKES
Sora's economics, in numbers. $15 million/day in compute against $2.1 million in lifetime revenue per public reporting. That's why the consumer app shut down. Worth remembering when you're picking which "freemium AI video" tools to build a workflow on.
NAB 2026 shipping report card. Of the major NAB 2026 announcements, Imagen Video is the first to actually be available the week NAB ended. Most of the rest are "coming soon," "Q3," or "later in 2026." If you build NAB-announcement-driven workflow plans, factor the lag.
DaVinci Resolve 21 is free. Public beta, full feature set, no studio license needed for personal projects. The free version of Resolve has been the best deal in editing for years; this version widens that lead.
ComfyUI is the platform layer for the open-source pipeline. 60,000 community nodes is the moat. If you haven't tried ComfyUI because it looked too technical, the institutional money rolling in will fund better documentation. The on-ramp gets easier from here.
Kling beta signups doubled in 10 hours on OpenArt. Demand is real. If you want to test the native 4K, claim your beta slot now — the queue is going to grow.
ONE MORE THING…
Video of the Week
GTA San Andreas — Movie Trailer | "Grove Street. Home."
Who made it: Solo (@Solopopsss on X) is an AI filmmaker with 120 million-plus views on his shorts and trailers. He's also the founder of Kickflix, a cinematic marketing agency that builds AI-driven brand films for crypto, consumer, and creator clients. Kickflix grew out of a project called KWEEN — Solo personally spent $20,000 in AI credits and 500 hours of work on those films before turning the lessons into an agency. He's the kind of working operator who knows what each tool actually costs in time and money, not what the marketing claims.
Why I'm sharing it: "Mostly" is doing the work. The interesting study for any creator is which 10-20% the maker did NOT let the model handle. Usually that's the close-ups, the eye contact, the dialogue beats. The seams are where the craft lives. Solo's work is a clean example of an AI filmmaker who makes those judgment calls rather than letting the model decide for him. Check out his Instagram profile
Worth a watch on a real screen.
FINAL THOUGHTS
This was a tools week that moved the floor up.
Kling 3.0 made native 4K from text a real option. ComfyUI made the open-source pipeline a $500M-valued platform with Netflix and Nike running it in production. DaVinci Resolve 21 turned what used to be VFX-house work into a slider in the program you already cut in. Imagen Video became the first NAB 2026 announcement to actually ship. And Sora — the product that defined the AI video conversation eighteen months ago — went dark, with the migration scattering across the field rather than landing on one replacement.
The pattern is the same as last week: the tools accelerate, the working creator picks the ones that fit the job, and the ones who pick well get more done with less. If you tested Kling 3.0 on something you actually planned to deliver, or if Imagen Video saved you an afternoon, write back. The practitioner data is more useful than the marketing footage.
Stay sharp. Keep creating.
— Larry
You don't need to be technical. Just informed.
Most AI newsletters are written for engineers. This one isn't.
The AI Report is read by 400,000+ executives, operators, and business leaders who want to know what's happening in AI — without wading through code, jargon, or hype.
Every weekday, we break down the AI stories that matter to your business: what's being deployed, what's actually working, and what it means for your team.
Free. 5 minutes. Straight to the point.
Join 400,000+ business leaders staying ahead of AI — without the technical overwhelm.
What did you think of today's newsletter?
AIography is the AI filmmaking newsletter for filmmakers, editors, and content creators navigating the biggest technological shift since digital.
If you have specific feedback or anything interesting you’d like to share, please let us know by replying to this email.
AIography may earn a commission for products purchased through some links in this newsletter. This doesn't affect our editorial independence or influence our recommendations—we're just keeping the AI lights on!
