I would love that. And I like interspersed between the random videos that these conversations can occur here. And yeah, inching toward the Hegelian dialectics—>Fukuyama “end of history” discussion I feel like…
There may be a cafe in Amsterdam in your near future. IBC.
It occurs to me that the future earth envisioned by Star Trek, is basically the fantasy of UBI for every individual on the planet. As a kid, I remember thinking, “Don’t they get bored??” Unless you’re Captain Kirk, frequently stimulated by hot alien chicks.
I read a meme the other day that said: The trillion dollar problem that AI is trying to solve is wages.
Been thinking about it ever since…
An interesting take on the art vs AI dynamic. This being flame community, it may not translate in direct way, but we are all hoomans here, I presume so there is that connectivity.
Steven Zapata Art - Is AI Winning the Battle for Social Media?
If you are working and want to listen to something in the background. He examines the shifting landscape for creators as generative media gains traction on major platforms.
This is a great point.
As I’ve come to understand very recently, it’s still the wrong question.
We don’t have to solve a wage problem, we have to solve productivity growth / pacing problem.
This is based on some conversations and experiments in another part of our company. And also crystalized in a WSJ story from this morning.
To frame it first WSJ: CEOs are split if they should just lay off workers, or if they should make them work more!
The initial reaction to this statement is: two abhorrent choices. But that is tainted by what ‘work more’ meant in the last two decades, particularly in the US market. We associate this ‘do more with less’ - or, we laid 10% off to redirect that cash to stock buybacks and you still have to do the same amount of work, just with fewer people. Horrible capitalism.
I’m now in the ‘work more’ camp. But it’s a different scenario. I’m about two make two people’s job redundant with AI, who can do it better and faster. But instead I’m re-deploying their work capacity to more meaningful and interesting work that enables the tasks AI has taken over. Thus we can keep the people, and everyone as a team is significantly more productive.
Now scale that across the economy. What if our existing workforce became twice as productive. We can suddenly do projects that were always beyond the roadmap cutoff, that were always too expensive for the budget we had, etc. That’s a good thing. But as companies that means we generate more output. Is there a market that can absorb this productivity gain? That’s likely to overwhelm the economy who will be unable to absorb an unprecedented and very steep productivity gain curve. So that becomes the problem to solve.
How can we pace and structure this productivity curve? But if we do so well, everyone can keep their job, their wages, and be happier yet.
In VFX terms, that means we can do more shots with the same budgets, or do more complex shots with the same budget. But do our clients have more business to give us? Some will, but scaling that productivity curve means matching our newly found excess capacity with our clients demand, who in turn has to do the same dance with their clients, who in turn will have to see if more people go to the movies or more people consume more marketing content in meaningful and profitable ways.
The answer is ‘yes, but in due time.’
You can draw similar pictures in editing - wrangling a documentary edit in the age of AI is a lot less work. Not automated, still human, but much more efficient. The story telling remains human, the big data challenge of doc editing gets AI help. Same with dialog editing, and any number of post production tasks.
If we come up short, and the economy cannot absorb the productivity gain, well, we can do the work in half the time, but will also just be paid for half time the time. Or we can collapse 2-3 people into one, and boot the other two. Lots of sad faces.
So lets figure this out and make it work.
We all will still be busy, just working differently (not more!). And I don’t think many will be lounging on the beach collecting UBI. That’s just a BS dream by the AI bros, so they don’t have to do the hard work of figuring this out.
Exciting times.
Is India the future of filmmaking?
Read Derrida’s Spectres of Marx. Fukuyama has been burnt by his “Hegelian” predictions of arrival/final Aufhebung, but still ploughing a form of liberalism. A very annoying “social philosopher” who loves sounding off (listen to him on Entitled Opinions for bim at his “height”
Yup… Many aspects of life are now being innovated outside the US where laws and lack of legacy allow fast adoption.
Sorry, but I just need to vent a little about AI. We’re currently working on a project where we have to create transitions from AI generated footage into live action footage. Long story short, it’s not working. The AI artists often don’t have a post production background and don’t understand exactly what’s needed to make this work. Not only does the image and camera movement change with every generation which doesn’t match the live action camera anyway but the resolution, length and FPS are constantly changing too. Then something weird happens in between that has to be fixed in post again, and so on. The list is very long. Nowadays, many shots are built in 3D or similar effects are created in After Effects, which works so much better.
For me, AI generated images and videos are a support for various things, but not a substitute just to save money so you don’t have to build everything in 3D or similar.
I’m not going to resist AI. It is very interesting and it looks really cool, but for now, I won’t be taking on projects that include AI in the future.
Having just gone through that exercise, I know exactly what you mean. In my case I was the AI artist. Some of these things can be controlled, some not. But for the ones that can, you need to setup a proper pipeline and discipline.
There are workflows where you can guide the camera. There are ways of writing prompts that declare certain un-targeted areas as ‘read-only’. If it matters, there ways of making this better. But you have to start with caring and knowing how and why.
To me that actually means it’s a prime opportunity for us to get involved in AI. Because we understand the workflow and the requirements. AI in post should really not be treated like an editor or colorist, which VFX folks often tend to ridicule (sometimes justified, but also often not), it should be treated as another effects discipline.
It comes down to having whoever does it understand what matters and that there incentives for bothering.
Don’t hire a YouTube kid to make your AI, or worse have the Creative Director DIY it, hire a VFX artist to do your AI. You’ll be happier for it.
The only reason we will use AI images, is because it allows to make things happen for less. No AI image is pretty enough to justify it’s existence if it weren’t for speed or money.
You’re not going to win this argument.
Did just that. In fact hired the VFX artist who said that. 
The only reason we will use AI images, is because it allows to make things happen for less. No AI image is pretty enough to justify it’s existence if it weren’t for speed or money.
You’re not going to win this argument.
I remember when similar argument was made for having ads on the channel. When people argued, what’s the harm, don’t use adblocker, you are in fact helping the artist and economy, and so on and so forth. I don’t think I need to explain to anyone who that argument aged like fine milk.
Same is the one with so called AI being made. And its going to age pretty badly too. Short term thinking and greed with no ability to learn from the past mistakes is how NFT’s were sold, than bitcoin and millions of other coins, now its AI etc.
But Have You Considered the Cost?
Fair argument. Don’t buy it entirely.
If your argument is applied to AI slop, absolutely. If your argument is applied at generating awful looking caricatures of what proper art direction may have yielded, absolutely. And that may well amount to 95% of all AI images (if you mix in the deep fakes and all the GIFs people try to be funny with).
But there is a small percentage of professionally made images that utilize AI that had proper art direction and all that, and which tell a proper story. But they couldn’t have been made with traditional techniques on the available budget or timeline.
For example, a shot in the video that is 98% identical to the original shot, but where one element was changed for story. That’s not an NFT or a hilarious GIF that ages in 5s. That’s something that shipped the job on budget and on schedule, with an enhanced toolset we didn’t have available a few years ago. Take a look back at Rufus’ shower gel video which he showed the BTS in great detail. We’ve seen many similar projects since. We’ve also seen projects where AI was used, where you wouldn’t tell if you didn’t know (because you weren’t supposed to notice).
It is funny though that the ai approach seems to be product over process, but so many audiences seem to be clamoring for process right now. Look at Project Hail Mary and how many behind the scenes videos you can find with their puppet and whacky sets. It’s 100% spectacle- but the spectacle is shifting since the days of Jurassic park, people, in my opinion seem to be looking for creative, novel, tangible and interesting processes to generate grand visions. I tend to feel that’s what a lot of people view as screen magic. But I don’t think for most people the gen AI process checks any of those boxes. It doesn’t mean it’s not hard work, or that it’s going anywhere or doesn’t require attention, just that there’s an extra hurdle here in my opinion.
Here’s an interesting question. In 2025 there was a Nobel prize in economics, that specifically focused on ‘creative destruction’.
It’s been quoted in an opinion piece about the demise of Spirit airlines WaPo (may be paywalled - also look in Apple News).
Relevant quote:
It’s tempting to consider Spriit’s demise as a failure requiring correction; in truth it’s evidence of the system working. […] Their central insight is that economic growth depends on the continuous cycle of innovation, transition and reallocation. Firms must be free to succeed and fail. Losses show resources are misallocated; profits show they were well used.
Which leads back to AI. If AI succeeds in super charging productivity, once available, its use frees up newly unnecessary cost.
There is of course the question of quality. And there is the question of human impact. Capitalism works out best for the people, if it comes with a decent slice of humanistic oversight.
If you look at fantastic shots VFX produces (whether CG or other common Flame workflows), they are immensely expensive to the point of cost prohibitive for all but big budget productions. Is that a misallocation of funds. It wasn’t in the past, because there were no alternatives. If they now do exists, does that fall under the umbrella of ‘creative destruction’?
To avoid jumping on a term or a quote from a news story without considering it wider, here is a more in-depth analysis if this an oversimplification, and how does the 2025 Nobel prize apply to the world of AI. It’s a pretty good read: Is that a convenient excuse or an important reality.
I think at a minimum it forces all of use to face the reality that budgets for our kind of work will get smaller, and we will have figure out how we can deliver fantastic shots to our clients ethically and economically sustainable for both us and them. In the Spirit of innovation and the art.
PS: That is for the consideration of AI as technology, that is independent of the behavior of the AI bros and the climate impact and behavior of CEOs cutting workforces. That’s a separate question, which Karma has on her roadmap.
1 Like
Gotta say and don’t want to say — I can’t argue that this short would have been better with human animators.