I hope I live to see the day when I can feed an AI the "Hamlet" skit from Last Action Hero, the original text of Hamlet itself, and the entirety of Arnold's work in the 80s and 90s and have it spit out the full length version for me to goon over and die happy.
My bet is 20 years for this to happen.
I'll take the other side of that bet and say we will see this happen in <5 years.
Agreed. This is moving faster than people not working in the field realize.
In fact, I'll wager to say that Disney and Pixar are both toast in just a few short years. You won't need Disney, $50-$100M in production funding, or more than one person to make a film. Films will cost $5,000. Then $1,000. Then very little at all.
Everything visual. Cartoons. Anime. Movies. Influencers. Actors. Porn stars.
Music is going to be right there with it.
And once the reviews are written by AI and the views are generated by bots, we can get those pesky humans out of the loop completely! Eventually we won't even need to have movies, just an agreement between two AIs that one will make a theoretical movie and the other will theoretically review it and watch it a million times. Heck, no human will have any money to go to the movies or buy any merch anyway.
The only humans left in Hollywood will be copyright lawyers!
I don't know why you have such a bad outlook about this.
I have wanted to direct epic fantasy films my entire life, and this is the first tangible shot I have at making something that might look like a "blockbuster". My chances of doing that without GenAI were slim to none. Hollywood is a steep pyramid.
I don't know how you can't be excited. This is the dreams-come-true timeline.
Hm. We may have a difference of outlook. I don't mind a fantasy blockbuster once in awhile, but to me the most interesting films involve interhuman drama that speaks to my own emotional experience, and often just consist of a lot of people talking with occasional plot twists and unexpected bursts of action. I went to film school for a couple years and dropped out, because as much as I enjoyed lighting and directing, I saw that everyone around me was obsessed with good effects and had no interest in good writing. [edit: That and it was insanely expensive and unjustifiable unless I wanted to be a commercial DP when I got out... and I wanted to be a writer.]
Yeah, I guess it would be awesome to just have an AI build your dream movie, but if you have a good script and a couple thousand bucks you can go out and find some actors and make something real right now. If it's good enough, you can get a shot at directing something massive where you actually have control over how your vision is executed, rather than sending your ideas into a "black box".
There's also the problem that visual fx are already pretty cheap, and no longer a reason to go see a movie. No one can even count the number of Viking shows or Lost Starship movies on streaming services, or tell them apart. It's not like when the original Jurassic Park came out and everyone was like I HAVE TO SEE THESE DINOSAURS. (My best friend recently described Avatar 2 as "that movie with the flying blue turds").
So my point is that the market is already oversaturated... just like mobile gaming. Chasing cool graphics, even if you get a monentary viral hit, you just end up lost in the flood of noise when everyone else has access to the same technology.
The pyramid is a good thing, actually, because it forces you to be imaginitive. The people at the top are less imaginative because they think their vast resources can compensate for bad writing. See: George Lucas.
Without a compelling story, a movie is worse than watching someone else play a video game. And with a compelling story, it doesn't matter if it's set on Mars or in your basement. I think it's also why the best games aren't the ones with the best rendering engines, but the ones with the best story and mechanics.
This sounds like the end-state is an increase in the relative value of extremely well-written films, as special effects and eye candy lose their luster as they become extremely common-place (see the Marvel universe of films, for an example of this happening already).
I can definitely see Netflix having four or five rows of AI movies above a ghetto section of "human-made" films. I already have to skip all the Marvel garbage. I mean, it's free to watch Twitch.
And, for that matter, the first few seasons of TNG.
Clerks was / is one of my favorite movies of all time.
(Far as Trek goes, I liked the original series.)
In theory yes, but imagine having to fight with another 8 billions directors to find someone who watches your movie. The day we can make our films at home (read as: everyone can make their films at home), is probably the day we'll need to create our own virtual audience as well, as the chances of being watched by some real person will be lower than winning a lottery.
Sounds kind of like YouTube but with better production value?
Afaik that arrangement works quite well by the standards of both the creatives and viewers (and presumably Google?)
that was my absurdist point about AIs making theoretical movies that would theoretically be watched by other AIs
Ah, I just clicked your profile and recognized that you're the person building the dream studio! Hah. It makes sense. And that's an amazing project idea. Maybe our views aren't entirely incompatible. I feel like an adjunct to your project would be a separate program / ML workflow that helped to hone screenplays and stories, tilting toward letting people express their experiences through these media in a way that broadened their ideas and expanded their potential audience before committing the ideas to images and pixels.
This reminds me of this quote from one of my all time favorite books, Dirk Gently's Holistic Detective Agency :
<quote> The Electric Monk was a labor-saving device, like a dishwasher or a video recorder. Dishwashers washed tedious dishes for you, thus saving you the bother of washing them yourself, video recorders watched tedious television for you, thus saving you the bother of looking at it yourself; Electric Monks believed things for you, thus saving you what was becoming an increasingly onerous task, that of believing all the things the world expected you to believe. </quote>
hehehheh. I just re-read that book for the first time in ages a couple of months ago, so Douglas Adams has probably influenced my thinking about these things ;)
Copyright lawyers are a great target for AI automation. I bet they’ll go first
I bet when CGI was first introduced (~40 years ago), you'd have said the the same or even more hyped: "Dude, in two years, all movies would be made on a computer. All 3D and VFX, by a single guy in his basement. No production or filming. Everything would be done on your computer, you don't need a crew. All these actors are gonna be jobless. Cameras? They're obselete."
Yeah, again, not how things works. Neither financially, nor the technology itself, nor the craft of film production. Not to be a party pooper, but the landscape will still look mostly the same in 5 years.
I agree to the extent that CGI took 40 years to fully mature to be accessible to the masses (read: see what Blender 4.0 can do on a kid's desktop[0]).
AI is maturing fast. It's already on some people's smartphones.
0: https://wiki.blender.org/wiki/Reference/Release_Notes/4.0/
But the stuff that works on people’s smartphones isn’t cinema quality. Which is precisely why it took decades for CGI to become accessible on our home PCs.
Generative AI is already powerful enough to replace big studios for short skits on YouTube. But a full length cinema-quality movie is a totally different ball game.
That all said, I definitely don’t think it will take 20 years to get there either. I do agree that the field is moving fast.
Given that %60 of a films production budget goes to marketing I'll say there will still be studios... They'll just have a higher margin.
If they have a higher margin there will be more studios competing for the same attention span/money and margins will come down.
I think live shows, improv and plays will stick around for a long time. But probably not for much money. It’s fun to do, it’s fun to watch.
Who knows?
So a fun thing is happening on Twitch right now! VTubers are improving [1] against LLM-powered actors. They're effectively able to put on one-person improv shows in fully dynamic worlds. It's wild.
Check out CodeMiko (probably NSFW). It's captivating to behold.
[1] (yes, we all know)
Media is popular and valuable not only because it's entertaining, but also because it's something you can talk about and relate to with others. So in that sense having the big brands with popular shows that most people watch is still relevant.
What does "few years" mean? Bet on a timeline or stfu.
And what makes you think that Pixar and Disney won't want to automate the process of movie creation? The movie streaming platforms all reached a equilibrium, so they would work on the "full self-making movie" I suppose?
The problem is that they are wealthy and politically powerful, so it's unlikely they'll go down without a fight.
I fully expect a lot of large-model stuff will become legally nuclear over time, much like the sampling culture in music died off.
Disney and Pixar have IP and parks, and a lot of cost with animators, actors, and directors to make new productions off of that IP. If in a few years all you need is 5000 Disney will make new IP for 5000 and structure it in the way that is most beneficial for their multiple franchises based off of their IP.
Disney won't be toast, they'll be a full continental breakfast in a 5-star hotel served by Star Wars Mickey in a waiter's costume.
I’ll take the other other side and bet that they’re in their room gooning right now.
Hey, that's a half truth!
These aren't AI generated, FYI. I thought so too at first but someone just put in a crapload of work collecting fake movies or TV shows which were in real movies or TV shows.
I think the idea is to use AI to turn these short snippets of imaginary shows into actual full-length show, not to generate more imaginary shows.
Check out Frinkiac[0]. An extremely well annotated dataset of all Simpsons stills for almost every single frame.
How has someone has not used this huge corpus of image and text to write their own Simpson episode?
0 :https://frinkiac.com/gifmaker/S07E04/1166047/1166047
Probably because "The Simpsons" has many (great) episodes already. The closest thing is probably "Dark Simpsons" or Simpsons Shitposts.
You might want to look up the current definition for "goon" It might not mean what you think it does
He clearly meant what he said.