The Stage 32 Lounge

Darrell Pennington yes! THE NAME OF THE ROSE 2

Debbie Croysdale it wasn’t me

Fran Tabor may we all live in interesting times

damn, damn, damn!

Then they died

Darrell Pennington Generative AI is working its way into studios as a tool, but it's far better than the retail stuff hobbyists are paying to play with. And it's got a long way to go - it can't replace the VFX pipeline because it cannot create simulation. IE. it paints a picture of what it thinks your description looks like, using a diffusion model. But it cannot create accurate shadows, fluid dynamics, raytracing, etc. because those are simulation effects and require lots of compute to do . AI just guesses. Also, it's about to hit a saturation wall - it will have consumed all available reference material shortly and mathematically then begin a trend to what is called "model collapse" unless some drastic changes in programming occurs, which in turn can require retraining. The other wall, the power wall, is for general AI, and is far, far more demanding than your post indicates. It's unsustainable. The alternative is quantum computing, which this year has proved viable. However.... if you think the AI industrial revolution is changing society, just wait for a couple years. Quantum computing is exponentially crazier... and doesn't require the same power.

Hi Shadow, thanks for the detailed response – you clearly know your stuff on the technical side. That said, I think there are some other angles worth considering:

On consumer AI tools – I'd push back gently here. Tools like Claude and ChatGPT have become genuinely transformative for a lot of professionals – writers, researchers, small business owners, production managers, smaller independent studios etc... The line between 'enterprise' and 'consumer' AI is blurring fast when someone can streamline real business workflows with a $20/month subscription. The value isn't just in the Hollywood VFX pipeline.

On simulation limitations – Fair point that diffusion models aren't physics engines today. But it's worth noting this is more of a priority gap than a fundamental limitation. Most AI development has focused on text, code, and image generation because that's where the perceived immediate demand was. Researchers are actively working on physics-informed AI models, and studios like Disney are already using machine learning to assist with environmental physics like snow and water simulation. The gap is closing; it just hasn't been the main focus yet.

On model collapse – This is definitely a legitimate concern the research community is watching. Worth noting though that mitigation strategies already exist – studies have shown that maintaining human-generated data alongside synthetic data prevents the collapse feedback loop. It's a challenge being actively worked on rather than an inevitable dead end.

On power consumption – The challenge is real, but some context helps: data centers account for less than 10% of projected global electricity demand growth through 2030 – EVs and air conditioning are actually bigger drivers. The grid strain is significant, but it's an infrastructure and investment challenge rather than a fundamental limit of the technology. Nuclear partnerships and efficiency improvements are already part of the industry response.

On quantum – The breakthroughs in 2024-2025 have been genuinely exciting, and the power efficiency comparisons are compelling for certain problem types. That said, quantum isn't quite positioned as a near-term replacement for classical AI workloads – you can't run an LLM on a quantum computer yet, and current systems can actually consume more power than classical computers for some tasks due to cooling and error correction overhead. Practical fault-tolerant systems are still targeting around 2029. More of a 'both/and' than an 'either/or' I think.

Appreciate you adding to the conversation – always good to dig into the nuances!

Shadow Dragu-Mihai, Esq., Ipg That is all heartening and provides me personally with glimmers of hope. I have no expertise in any of the technology that AI rests on or the actual challenges that the technology will continue to face as it tries to move forward but in general I do think that we collectively can use history as a guidepost for potential outcomes and my concern is that the level of resources that will continue to be funneled to continue to improve AI and push it ever closer to universal adoption in all aspects of life. So I hope the pessimism many have is warranted and I am Chicken Little as it relates to what I think are likely negative outcomes. But beyond all of that, I think there is a malevolence being crafted by just a handful of people that are going to provide us all the options we have in life. All in the name of ease, comfort and progress.....

Darrell Pennington On S32 and in our circles we talk a lot about generative AI, but that's really small AI. Big AI is the real game changer and it's already a fact. We're in an industrial revolution and there's no going back. Whole industries are being re-configured or going the way of the dodo. Economies are shifting. Consider that ~20 years ago we in North America started farming out tech support and customer service support to the third world. We lost a lot of jobs here - to the advantage of people outside the country. However, those countries have not advanced much, while our own work force has adjusted. Now one of the things AI is wiping out almost entirely is tech support and customer service. Those economies that have relied on that as a product are pretty screwed. Lower to mid level programming, website design, and lots of research positions (at least the internet research kind) are no longer needed as well. I myself have been able to forego a fully paid researcher for our AI and quantum computing initiative, and I get my results in minutes instead of weeks. This, as well as the fact that AI is now heading to a spot where it decides what objective truth is... and you can see the challenges. But it's far too late to complain - the trajectory started in WW2.

Joshua Young I do not disagree on most of what you say, though as far as image/video I must stand my ground. For several reasons but one being that script writing software never made anyone into a writer who wasn't already using a pen and pencil, and the hi-rez cell phone camera never made anyone a filmmaker who wasn't already shooting something on their camcorder, and more recently Unreal Engine hasn't given anyone a lift into filmmaking who wasn't already doing VFX. Generative AI is a visual tool. I see a host of hobbyists playing with the new toy, but I don't see the lines between professionals and hobbyists converging in reality. The tools are just now starting to be useful to professionals and it isn't a magic formula. In fact, it can create almost as many problems as it solves, including a copyright minefield and the "uncanny valley" effect, which has been clinically proved on AI generated video. On simulations - no. Generative AI will not be able to do this, until it merges with actual simulation techniques, raytracing, Monte Carlo processes, et al. which are all costly in compute. We are working on that integration ourselves, but it's not going to be something that any consumer can take advantage or, or should care about, IMO

Thanks for making this post, Darrell Pennington! Happy Holidays! Hope you and your family enjoy them and hope you all have an incredible new year!

Elle Bolan You are right! All stories have already been told. In the metamodern world, any plot will inevitably refer to something else. I am a big fan of sci-fi, so "military experiment" remindes me Marvels Captain America and Winter Soldier. For me the question is if superhero films are still actual? I feel, that we are in the era of "Everything Everywhere All at Once" and other "new sincerity" A24 films

Xyphalavino Muthii Thank you for sharing. Your story intrigued me, but some points remain unclear to me.

1) Who is the hero

The hero is clearly defined: a young man who has survived an illegal military experiment.

2) The hero’s goal

The hero’s goal is not stated explicitly; now I can only assume it to be survival. Why are they hunting him — to kill him, to strip him of his free will to turn him into a true weapon? In The Bourne Ultimatum, for example, the fleeing protagonist’s goal is to uncover the truth about his origins and permanently destroy the program that turned him into a killer.

3) The antagonist

The main antagonist is clear: the organization that created the experiment and is now pursuing the hero.

4) The outcome — win or lose

This is not always stated in a logline, but it is important that you know it. Even in simple cat-and-mouse stories like Tom and Jerry, it is clear for creators who ultimately wins and how. So it helps to build a plot. In your case, like all superhero films, I assume there will be a happy ending. Which is always boring for me, but the general audience expects and appreciate it.

Xyphalavino Muthii I have a question about originality. The theme of a person bonded with artificial intelligence and gaining superpowers appears in other projects — there are many films and series about AI and enhanced humans — but the exact combination of nano-AI + super speed + a military organization does not seem completely standard; the elements are familiar, but the combination could feel fresh.

Whether studios and audiences need another super-speed hero after the X-Men films, and how character rights are protected, I don’t know. What is important is to write unique plot twists and sincere emotional character arcs that resonate in the era of metamodernism.

If I were you, I would research the superhero movie market from 2019 to 2025 — do these films still perform at the box office today, and what do experts forecast regarding any revival of the genre. In my opinion, even James Gunn hasn’t saved Superman yet — although films like The Suicide Squad (2021) and the Peacemaker series seem interesting. But that is my personal view; the real question is what the numbers say.

Antony Voronov Elle Bolan I get it and i know all stories have been told but what makes a character special isn't whether he existed or not but it's how you tell the same story in a very different way.Like for example if you look at action movies, you'll find them having something in common like the protagonist having to diffuse a bomb in the last minutes,so many movies have this but whenever you watch these movies it feels new because it's a new a character having the same story but now told in a different way. So coming back to my character PULSE,i did my research on other existing speedster and made sure i gave him a very different story than the ones before making it a new story even though the idea of speedsters already existed.

I get all of that too.

My comment was actually positive. These types of stories have seen repeated success through 50+ years. A fresh, timely version of a similar story is a good thing. Not much is unique, we make it unique in delivery.

Which I believe your project does quite well.

Merry Holiday Season, Juliana Philippi! "I have wanted to sleep more, but more has been happening!" I relate. I was gonna write on Christmas, but I'm taking a break after today.

Congratulations on believing in yourself and your career taking off! Thanks for the support and positivity you bring to the community. Hope you and your family have a wonderful time on the holidays!

Maurice Vaughan Thank you so so much!!!!! Thank you for your support, for your insight, your kindness, your willingness to be so open and caring, and honestly, interested in helping all those who are here : ) The best to you, and yours, the best holidays, and a wonderful new year!!!

You're welcome, Juliana Philippi. Belief in ourselves can help us write scripts, pitch, network, and more. Thanks.

Happy holidays to you.

May the new year be full of success, the realization of your dreams, love, happiness, and everything you hope for. I just saw on your account that you’ve won 18 awards—that’s incredible. Congratulations, and may you achieve even more next year.