Rich Klein✓

I have a super exciting update coming soon to AI Film History! 🔥 A new way to learn and experience the history. Hoping to release it this week.

It’s that time of the year again, when I review all of my active creative tools subscriptions for the upcoming year before all the big sales. Things shifted a lot this year, so the decisions are getting tough. Which tools are on your must-keep list and which ones are on the chopping block? Right now, my must keep list is: - Midjourney: still no match for its creativity, exploration, and overall fun value (is that a thing?) - Runway: the all-in-one powerhouse generative suite with great MCP and API tools. - ElevenLabs: works so well with agents via MCP with Hyperframes. Quality in V4 speech model is still unmatched. - MeshyAI: huge boost for 3D work and web gamedev. Also works great via MCP via Claude code. I won’t post my drop list out of respect for the companies. But these are tough decisions, guided by my usage and overall value.

This week (well, two weeks) in AI film history 🎞️ 🎃 1. Oct 29, 2021: @Somnai_dreams releases Disco Diffusion, building on prior work to pair diffusion models with CLIP for text-guided images and animation. It became the training ground for a generation of AI artists. 2. Oct 23, 2023: Martin Haerlin makes "Insurance covers everything everywhere all at once." for Runway's first Gen:48. 3. Oct 31, 2024: @BLVCKLIGHTai releases Whispers in the Dark, a 30-episode cosmic horror series cut into a 1h 43m feature, at a time when most generative work was still 10 second clips. 4. Oct 23, 2025: Lightricks (@ltx_io) announces LTX-2, the first open model to generate picture and sound in a single pass, in native 4K on a 12 GB consumer GPU. The open weights were released in January. 5. Oct 25, 2025: @brianjamesgage releases Pumpkin Patch, a pumpkin-headed vigilante breaking into the prison colony District 13 to rescue the President of Halloween. Its sequel, Moonbreaker, was released this month. 6. Oct 29, 2025: @MetaPuppet's KYRA, a sci-fi short with polished visuals, coherent characters and atmospheric world-building from a solo creator. 7. Oct 31, 2025: exactly a year after the first, @BLVCKLIGHTai releases the 84-minute sequel, Whispers in the Dark 2: Digital Purgatory. Links on the timeline 👇. Enjoy!

1. 2. 3. 4. 5. 6. 7.

I was grateful to attend the Lumara Film Festival hosted by @AskVenice. As someone who's watched a lot of generative films for my AI Film History project, the biggest takeaway from tonight is that we've reached a new level of quality and storytelling. I first noticed the shift in June at Runway AIF, and tonight deepened my belief. You first realize the shift in those magical moments when you're carried away into a story. Where the distance between you and the screen feels nonexistent. The finalists screened today reached that level. These are true artists excelling at their craft. 👏

This is a big release! 🔥 All-in-one, fully local, generative filmmaking suite.

Big Announcement: Meet Imagination Studio🎬🎵 by @u1tra_instinct and Me! Our local open source all in one creative platform on your own hardware has been fully redesigned and rebuilt from the ground up. Video,Audio, Music, Image. Releasing ASAP with latest+best local models.

Just pushed some nice updates to the Artist and Film pages on AI Film History(.org). The link is in my profile. There are new screening modes for films and more curated/showcase films to watch. I also added some of my favorite films from each era to a new "Essentials" section. If you're just getting started with AI Filmmaking, this is a great place to learn about the history of the field and how the films have evolved from both a technical and artistic perspective. If you have any other suggestions, let me know.

Can you tell a story with enamel pins? Share yours. "Unexpected friendships" ☺️

Excited to share some big updates to the AI Film History site. 🎞️ The goal of these updates was to add different ways to engage with the site as the timeline grows. Most viewers will rarely read the entire timeline, so I added new sections to highlight films and also spotlight different artists each week. There's also a new "Watch the history" view that autoplays the timeline in screening view. Just sit back, relax and watch clips from every era in about 11 minutes. Some other big updates: 1. You can now pick up where you left off on the timeline. 2. I split 2026 into two eras (it was way too big): The Shakeout (early 2026) and Prime Time (late 2026). 3. The Workflows page was overwhelming, so I reviewed all the info (again) and rebuilt it from scratch (again 😀). It now focuses on the evolution of how AI filmmakers gained more control over shots and the different tools, models, and techniques used to get there. It's a much more approachable read now. Link below: 👇

Learn more and share your films at

Do you prefer your “For You” feed to be filled with varied interests or focused on your primary work and communities? I prefer it varied then using lists for more focus, but the “For You” algorithm seems to trend towards mostly focused mixed with popular posts.

This week in AI film history 🎞️ Check the links below for details and to watch the films. 1. Oct 14, 2024: @AdobeVideo launches the Firefly Video Model at Adobe MAX, with a commercially safe pitch, camera controls and Generative Extend, which adds frames to a clip right inside Premiere Pro. 2. Oct 14, 2025: @GoogleDeepMind ships Veo 3.1, with richer audio, native 1080p, clips up to 60 seconds and editing inside Google Flow, turning the model into a full production environment. 3. Oct 15, 2023: @AIandDesign releases Cyberpunk Dreams, a near-three-minute music video made from Midjourney stills animated in Runway Gen-2 and cut in CapCut. 4. Oct 16, 2025: Eyes On You, a Twilight Zone-style psychological horror short made with Veo 3: a man's ordinary movie night, and eyes on screen that might not be acting. 5. Oct 17, 2023: Martin Haerlin makes a 39-second BMW ad from six tools in a row: Midjourney, Runway, Pika, ElevenLabs, Tracksy and Topaz. One of the earliest AI car commercials. 6. Oct 17, 2024: Tom Paton's Where the Robots Grow, an 88-minute AI feature from AiMation Studios, about robots farming a new world and the last human baby.

Read more on each 👇 1. 2. 3. 4. 5. 6.

A Magical Evening!

Some eye-opening numbers in this article by @AnthropicAI. If you've been working in software development for your entire career (like me), it's hard to wrap your head around this scale. The SDLC structures we all grew up on are breaking down quickly. This article provides a glimpse at new structures for a company working at agent scale.

New on our Engineering blog: Creating in-house source control. In May, we merged around 2,500 code changes per day. In September, we merged more than 30,000 per day. As a result, we became bottlenecked by developer infrastructure built for human (rather than agent) scale. Here, we share how we built a new version control system—and fully migrated our monorepo from git—in five months:

Great time- and credit-saving tip from Tom! 🔥 We've all been there. Seedance 2.5 generates a perfect clip, but added random background music. This technique, using Meta's SAM model, helps remove unwanted music while keeping the dialogue and audio you want.

Useful tip for anyone using SeeDance 2.5 It's easy to remove unwanted (and unprompted) music from a video with Met's SAM Audio. You can rescue a generation by isolating any sound, including music and removing it. This allows you to hold onto Sound FX but add your own soundtrack which will probably sound better and will add consistency across generations.

Another funny one by the Captain! 🤣 Great series.

They say age is just a number. Some don't like getting older personally I see it as an achievement of survival.

Lately, I've been more focused on building apps and games for myself or to share with family and friends, without worrying about whether they can be monetized. Perhaps this is the future of smaller-scale consumer application development. No longer a profitable business, but more personalized and user-driven. Before coding agents, creating real solutions to most problems required significant time and money, which wasn't realistic for personal side projects.

Opus 5.5 continues to surprise me. This morning I thought I was done with a small game project, then I used this simple prompt in Claude Code: "Do you have any other ideas to make the gameplay smoother and/or more engaging?" The suggestions from Opus 5.5 were incredible. It turned those suggestions into a phased plan and started working. 🤯 I am going to use a similar approach for my video and web development projects as well.

Would you play this Groundhog Sim game?