Your weekly skim of what the Machine Cinema crew debated, discovered, and dropped into the chat. Too busy to read? Have a listen instead.
Edited by Elizabeth Kealoha. Images and Pod Digest by Ant Neely.
Want to know what 1,000+ creators are whispering about behind closed doors? Consider yourself invited.
Overheard in “Basecamp”…
Full disclosure, we had our robot friend help us pull all this together and sometimes they are prone to making harmless mistakes.
🤖 RoboDojo Launches
A new weekly RoboDojo session and dedicated chat group kicked off—designed as a hub for tool updates, workflow hacks, and AI news sharing.
Weekly sessions to cover tool drops + best practices
Simultaneous WhatsApp group for ongoing discussion
💬 “Feel free to post latest news, announcements, ask questions, share workflows.”
🎥 Fan-Made Witcher IV Trailer
A community team dropped a slick Witcher IV concept trailer, earning praise for fluid camera work and standout visuals.
Notable use of AI compositing for cinematic moments
Discussion of VEO3’s strengths/weaknesses in animation styles and lip-sync
💬 “The standards are high!! Congrats to the artist.”
🧰 Workflow Debates: Consistency, Styles & Tools
Filmmakers swapped notes on tool selection for specific styles, lip-sync accuracy, and animation.
VEO3 praised for flexibility but flagged for morphing faces
HeyGen, Hedra AI strong for lip-sync
Boords + CapCut recommended for storyboarding and editing
💬 “I left as Tom Cruise in Minority Report and came back as Neolithic Man.”
🎭 Dor Brothers Awards Debate
The launch of the Dora Awards sparked heated critique around ethics, labor practices, and judge credibility.
Concerns about Fiverr labor pipelines and exploitative models
Questions on “anti-censorship” rhetoric vs. fair compensation
Debate over marketing-driven judge selection (influencers vs. experts)
💬 “The creative community deserves better than awards ceremonies that exploit artists while using political positioning to build personal brands.”
🐉 Luma & Seedream Impress (But Not Perfect)
New drops from Luma, Seedream, and Reve wowed many with cinematic quality—though flaws remain.
Dragon shot from Luma demo called “amazing”
Reve praised for thoughtful UI/UX
Concerns over morphs, repetitive patterns, and “habituation” in agentic video
💬 “Custom styling falls apart at the agentic level—you end up with whitewashed videos.”
⚖️ Copyright, Indemnification & Insurance
Big focus on client concerns: which tools are legally safe?
Midjourney often banned in commercial use; Luma, Veo3, and Runway preferred
Indemnification usually locked behind enterprise plans ($$$)
Insurance for AI creative liability seen as inevitable
💬 “You could get a logo scrawled on a shirt that infringes without you realizing it.”
🗓️ AI Production Scheduling & Costing
Producers asked: how many AI shots per hour is realistic?
Estimates ranged 1–2 quality images/videos per hour
Scene complexity and subject matter drastically affect timing
Proposal floated for a “budgeting cheat sheet” or even a class on AI project scheduling
💬 “Maybe Machine Cinema will sponsor a schedule jam—only cost is lunch.”
🧩 Creative Provenance & IP
A lawyer’s advice spurred a discussion on documenting AI/human collaboration.
Suggested methods: timestamped transcripts, seed prompt records, Google Meet recordings
Key: capture human-origin “creative seeds” before machine iterations
💬 “Those data packs are the gold.”
📚 Everything Is a Remix (Still)
Myth, IP, and remixing got revisited with nods to Campbell, Rowling, and Star Wars/Harry Potter overlaps.
Reminder that originality is often about recombination and subtle differentiation
“Perfecting the wheel” as valuable as inventing it
💬 “Everything is a remix, baby.”
🔥 New Tools & Workflows
Navigating the Rapid Pace of Innovation The group's ongoing discussion of tools saw several new names emerge as serious contenders. Members shared their favorite models and workflows for various tasks, highlighting the rapid pace of development in the space.
Image Generation: Members praised Reve for its intuitive interface and user flow.
Video Generation: Veo 3 and Luma were cited as top tools for video-to-video work, with Luma's new HDR and cinematic capabilities generating excitement.
Character Consistency: The group debated the necessity of Loras post-Nano Banana and Seedream, with many finding that single-image character hold tools are now sufficient.
High-Res Output: Magnific was recommended as the go-to tool for generating high-resolution images for large-format printing.
Real Creative “Pick of the Week”
Each week the Machine Cinema members + the Real Creative team obsess over social feeds in search of the world's best AI creative video, gaming, and music projects.
This week one project stuck out: "The Adventures of Neemo Green" from Neural Viz, who continues to create some of the most inventive AI films out there. The 11 minute short is a super clever sci-fi adventure which took 120 hours to make by just one person and used tools Nano Banana, Veo 3, Kling, Runway, Elevenlabs, Sora, Udio, Suno and Photoshop. In addition to stellar use of AI, it's a great story with excellent dialogue and you can find more from Neuralviz here: YT, IG, Merch, Patreon.
You can find this and over 300 AI filmmakers and their projects over at realcreative.ai which features some familiar faces from the Machine Cinema community.
Overhead in “We Love Robots”:
Here’s what bubbled up in the wider creative/tech landscape.
Full disclosure, we had our robot friend help us pull all this together and sometimes they are prone to making harmless mistakes.
🌍 Startup–Enterprise AI in Tokyo (Demo Slots Open)
A partner call went out for an AI showcase in Tokyo on Sept 26 (co-organized with NVIDIA & FPT Japan). Startups can demo to ~200 attendees across enterprise, government, and CVC.
Opportunity: Live enterprise exposure + warm intros
Action: Apply via Luma; organizers are curating demos
🧰 New AI Device Buzz (Whispers of an ~$800 Price Point)
A late-night flurry over a teased AI gadget sparked debate on whether “smartphone-level” pricing will stall mainstream adoption—or kick off a new category entirely.
Signal: Early adopters are in even at premium pricing
Watch: Whether v1 utility beats “novelty device” stigma
💬 “I’ll be getting one personally but really excited for this space to evolve.”
🧷 Can AI Code “Real Work” Yet?
A shared chart framed current capability: strong on narrow coding tasks, promising on multi-hour pro tasks soon, uncertain on open-ended architecture/judgment.
Near term: Pair-programmer, boilerplate, unit tests = ✓
Hard part: Product sense, system design, tradeoffs = TBD
💬 “Thanks, ChatGPT!”
⚖️ Studios vs. China’s Image Models (Lawsuits Heat Up)
Members tracked news of Hollywood studios suing Chinese AI firms over training/use of copyrighted material, reopening the “what’s fair use for training?” fight.
Risk: Fragmenting global precedent across jurisdictions
Outcome to watch: Carve-outs vs. licensing schemes vs. bans
🧩 Chip Chess: China, NVIDIA, Intel—And Energy
A dense back-and-forth on chip restrictions, China’s potential to scale “good-enough” silicon + massive data centers, and the ASML lithography bottleneck.
Thesis: US = energy constraints; China = chip complexity constraints
Counter: Partnerships (e.g., NVIDIA–Intel) keep moving pieces
💬 “Proper gloves off state-level space race for superintelligence now, isn’t it?”
🛠 The Browser Wars (Again), Now with AI
With Google announcing new AI features for Chrome, members called the browser “the next OS” in the AI era.
Stakes: On-device models, agentic autofill, live UI rewrites
Question: Who owns the user’s action history & data trails?
🔗 Quick Hits & Bonus Links
🧪 Tools & Industry
OCME: Shaping the Future of Content Commercialization and Distribution
Benchmarking Creative AI- Creative Arena
TV Academy Establishes AI Guidelines for Members
AI Liability- The Next Step
Copyrighting Solution for AI-Generated Content
📚 Business & Economics
Runway launches Foom!
Higgsfield Venture Program
Comfy(UI) in the Cloud
ElevenLabs introduces Studio 3.0
🎥 Film, Art & Festivals
The PGA Innovation Award 2026
House of Prompt: Open Call for AI Artists
Leonardo AI Imagination Fund
BLVCKL!GHT announces upcoming project
📢 Community & Events
Synthetic Narratives Event
GenTalks Community Call September 17, 2025
Our GenDojo Community Call is a weekly digital get together to connect on how we are all making our way through this new era for creative industries and AI among other emergent technologies. Each week we invite artists, builders, thought leaders to share their knowledge, their works in progress and their ideas on this emergent space.
If you’d like to be added to the recurring invite please DM.
Session Overview
This week’s GenTalk featured two deep-dive guests:
Kelsey Farish – London-based lawyer specializing in AI, media, and creative rights.
Erik Weaver – Head of Virtual & Adaptive Production at USC’s Entertainment Technology Center (ETC).
The conversation spanned legal frameworks for AI in creative work and cutting-edge pipelines for AI-driven studio filmmaking. 📺 Watch Here!!
Highlights
1. Kelsey Farish – Legal Frontlines in AI & Creativity
Career arc: Among the first lawyers peer-reviewed on AI + creative industries (since 2018).
Core expertise: Digital replicas, voice cloning, consent contracts.
Case studies:
Licensing an actor’s voice for an X/Grok chatbot (balancing scope of license, confidentiality, psychological impact).
Handling requests to voice-clone a person with dementia—complex legal/ethical territory.
Key messages:
Consent ≠ informed consent → explain exactly how likeness/voice will be used.
Contracts should be in plain English → a contract is a story.
Do a “vibe check” → align AI use with your values & audience trust.
Don’t use ChatGPT as your lawyer (“you’ll look like a jerk”).
Practical tips:
Define an AI philosophy (guiding principles, even just keywords).
Vet T&Cs carefully (scope of license, retraining clauses, branding rights).
Trust but verify → educate yourself on how the tools actually work.
Lawyer up (doesn’t need to be prohibitively expensive).
2. Erik Weaver & ETC – Studio-Grade AI Filmmaking
Link to his presentation.
ETC’s mandate: Founded by George Lucas post Beta/VHS war → align studios on digital cinema standards. Weaver reports directly to CTOs at Disney, WB, Paramount, Sony, Epic, etc.
ETC’s AI pipeline:
Projects: Chickaboo, The Bends, Pathways.
Uses GenJam-style teams mixing AI artists + Hollywood veterans (DPs, producers).
Heavy infrastructure: ShotGrid integration, Dockerized environments, AWS + Seagate storage, RunPod bursts for training, Topaz upscaling.
Workflow experiments:
Blobfish character: licensed Oceanic Institute photos → 3D conversion → LoRA training → 4K EXR outputs.
Pathways: live-action iPhone shoots (sensor-rich) → AI background swaps, Blender comps, Runway first-frame, BBL.ai relighting.
Standards focus:
Aiming for 4K, 32-bit EXR, film scans to homogenize look.
Creative intent mapping → comfy nodes to preserve authorship trail for IP/copyright.
Union questions:
ETC follows SAG rules (e.g., delete digital doubles at end of production).
Little direct dialogue with DGA/IATSE yet, but Weaver sees ETC as a neutral lab to help set realistic standards.
Budgets: The Bends = ~$20k (mostly volunteer time), compared to $100k+ traditional VP shoots.
Volunteers wanted: ETC is actively seeking AI artists & compositors (Nuke/AE).
Key takeaways:
“This is about storytelling, not tech.”
GenAI is “faster VFX,” still requiring significant human labor.
The real horizon: new models for Hollywood production (not 1-person, not 100-person, but something in between).
Hot Quotes
Kelsey: “A contract is a story. If you don’t understand it, that’s your lawyer’s fault, not yours.”
Kelsey: “You will look like a jerk and an idiot if you use ChatGPT as your lawyer.”
Erik: “If you think this is about technology, you’re confused. This is about story, plain and simple.”
Jena Bodell (ETC): “This is just faster VFX—it still takes a lot of human manpower.”