I haven't seen comments regarding a big factor here:
It seems like OpenAI is trying to turn Sora into a social network - TikTok but AI.
The webapp is heavily geared towards consumption, with a feed as the entry point, liking and commenting for posts, and user profiles having a prominent role.
The creation aspect seems about as important as on Instagram, TikTok etc - easily available, but not the primary focus.
Generated videos are very short, with minimal controls. The only selectable option is picking between landscape and portrait mode.
There is no mention or attempt to move towards long form videos, storylines, advanced editing/controls/etc, like others in this space (eg Google Flow).
Seems like they want to turn this into AITok.
Edit: regarding accurate physics ... check out these two videos below...
To be fair, Veo fails miserably with those prompts also.
Are users of the $20 tier really going to have to deal with that obnoxious bouncing watermark I wonder? The previous watermark could be cropped, but I often didn't feel the need to as I use it for fun, but that would make me not want to show anyone.
That seems like an awful use of technology like this. I would imagine they mean to use that for serving ads, but how do you even generate conversations with ai slop plus product placements? I could see it working sometimes but I doubt it scales.
I wonder if they're going to license this to brands for heavily personalized advertisement. Imagine being able to see videos of yourself wearing clothes you're buying online before you actually place the order, instead of viewing them on a model.
If they got the generation "live" enough, imagine walking past a mirror in a department store and seeing yourself in different clothes.
We probably haven't even scratched the surface of what will be done with this tech. When video becomes "easy", "quick", "affordable", and "automatable" (something never before possible on any of those dimensions) - it enables countless new things to be done.
When the dust settles , that's probably going to be the most common application of these video models. Making automated social content kind of defeats the purpose; people empathize with other people, not with AI . (I guess that's why they didn't also make their interview video via AI)
But Sora /VEO will probably also revolutionize movies and tv content
its called Virtual Try On (VTO) and there are plenty of models going there for static gfx, it is very reasonable to expect soon emerge those for video VTO.
Accurate virtual try on however is quite difficult, and users will quickly learn to distrust platforms that just generate something that"looks right".
You can prompt with a normal size 8 dress and "kim jungle un wearing a dress" and it will show you something that doesn't help you understand whether that dress would fit or not. You can ask for a tube dress and it will usually give him a big bust to hold it up. It's not useful for the purpose of visualing fit.
It will definitely be used for such just like image models already are for cheap tenu clothes, and our onions shopping experience will get worse.
Maybe this needs purpose built models like vibe-net or maybe you cab train a general purpose model to do it, but if they were spending the effort necessary to do so they'd be calling it out.
You don't need generative AI for that at all, snapchat filters have existed for a decade and are the same concept. A lot of brands have already adopted that.
What makes TikTok fun is seeing actual people do crazy stuff. Sora 2 could synthesize someone hitting five full-court shots in a row, but it wouldn’t be inspiring or engaging. How will this be different than music-generating AI like Suno, which doesn't have widespread adoption despite incredible capabilities?
> Concerns about doomscrolling, addiction, isolation, and RL-sloptimized feeds are top of mind—here is what we are doing about it.
> We are giving users the tools and optionality to be in control of what they see on the feed. Using OpenAI's existing large language models, we have developed a new class of recommender algorithms that can be instructed through natural language. We also have built-in mechanisms to periodically poll users on their wellbeing and proactively give them the option to adjust their feed.
So, nothing?
I can see this being generated and then reposted to TikTok, Meta, etc for likes and engagement.
The most interesting thing by far is the ability to include video clips of people and products as a part of the prompt and then create a realistic video with that metadata. On the technical side, I'm guessing they've just trained the model to conditionally generate videos based on predetermined characters -- it's likely more of a data innovation than anything architectural. However, as a user, the feature is very cool and will likely make Sora 2 very useful commercially.
However, I still don't see how OpenAI beats Google in video generation. As this was likely a data innovation, Google can replicate and improve this with their ownership of YouTube. I'd be surprised if they didn't already have something like this internally.
> the ability to include video clips of people and products as a part of the prompt and then create a realistic video with that
This is something I would not like to see, I prefer product videos to be real, I am taking a risk with my money. If the product has hallucinated or unrealistic depiction it would be a kind of fraud.
The main lesson I learned from the March ChatGPT image generation launch - which signed up 100 million new users in the first week - is that people love being able to generate images of their friends and family (and pets).
I expect the "cameo" feature is an attempt at capturing that viral magic a second time.
> The Sora 2 system card claims Sora can resist generations of "political persuasion".
To actually do that, it would need to have the evolving contextual knowledge of current events and the reasoning power to be able to identify prompts which, when requested, would likely have use in political persuasion, which would be a bigger breakthrough in AI than anything they are promoting.
Conclusion: it can’t actually meaningfully do that, though it will probably reject some subset of prompts involving topics pre-identified to it as politically sensitive.
This is a good point. It is hard not to see this tiktok video of an unblinking man smiling at a soft focus sasquatch and think of Marconi’s first broadcast across the Atlantic. I challenge any man to look at this unsettling video of a dog astronaut and not compare it in his mind to the first integrated circuit, or alternating current.
Sheeeeeeeeeeesh. That was so impressive. I had to go back to the start and confirm it said "Everything you're about to see is Sora 2" when I saw Sam do that intro. I thought there was a prologue that was native film before getting to the generated content.
I'm sorry but that's a gross exageration. If any of this was real film then I'd start a gofundme page for OpenAI to get better video production equipment and team because that would be laughably bad.
If anything, it looks a lot worse than a lot of AI-generated videos I've seen in the past, despite being a tech demo with carefully curated shots. Veo 3 just blows this out of the water for example.
It's not an exaggeration to me? I literally stopped the video and went back to the start and re-read. You're more than welcome to speak about your opinions and experiences, but I'm speaking about mine.
I'm over here thinking, "It felt like just yesterday I was laughing at trippy, incoherent videos of Will Smith eating spaghetti."
I love the progress we're making. I love the competition between big companies trying to make the most appealing product demos. I love not knowing what the tech world is going to look like in six months. I love not thinking, "Man. The Internet was a cool invention to have grown up in, but now all tech is mundane and extractive." Every time I see AI progress I'm filled with childlike wonder that I thought was gone for good.
I don't know if this represent SOTA for video generation. I don't care. In that moment I found it impressive and was commenting specifically on the joy I experienced watching the video. I find it frustrating to have that joy met with such negativity.
I fully understand the hype but the initial scene with Sam feels nothing like how any self respecting video producer would create. The jump cuts mid-sentence are extremely jarring, certainly not framed in any traditional sense, and he's almost entirely out of focus.
Points though for the completely expressionless line delivery, it completely nailed that.
Don't worry. AI is going to be monetized and extractive in no time. Just like Social Media went from "fresh, fun and cool new tech" to "how did we let this horrible beast take hold of the world," AI will take the same path. In 10 years or sooner, when 99.99% of what you read, hear, and watch is AI slop, you're going to post "This used to be a cool invention!" if there's even a place left for humans to post by that time.
I agree. It will absolutely get there. Such is the trend of all scientific inventions. A breakthroughs occurs, prosperity follows in response, hedonic adaption causes satisfaction to regress to the mean, and then people squeeze every remaining drop of value out of the technology while we wait for those capable of true innovation to work their magic once more. I don't find it idyllic, but I accept it as the way the world works. It feels like a force of nature to me.
The period we're in is fleeting. I think it should be acknowledged and treasured for what it is rather than viewed with disdain because of what is inevitably to come. I stopped using Facebook and never moved to Insta/TikTok when things began to feel too extractive, but, for a good decade there, I felt so close to so many more people than I ever thought possible. It was a really nice experience that I no longer get to have. I'm not mad at social media. I'm happy I got to experience that window of time.
Right now I'm very happy to be using LLMs without feeling like I'm being preyed upon. I love that programming feels fresh and new to me after 15 years. I'm looking forward to having my ability to self-express magnified ten-fold by leveraging generative audio/visuals, and I look forward to future breakthroughs that occur once all these inventions become glorified ad-delivery mechanisms.
None of this seems bad to me. Innovation and technological progress is responsible for every creature comfort I have experienced in my entire life. People deserve to make livings off of those things even if they weren't solely responsible for the innovation.
I get what you're saying. I see it too. That said, a lot of people won't notice the flaws, especially with these fast, choppy cuts. By the time you realize the neck is way too long or whatever, its 2 cuts later.
The example prompt "intense anime battle between a boy with a sword made of blue fire and an evil demon demon" is super clearly just replicating Blue Exorcist https://en.m.wikipedia.org/wiki/Blue_Exorcist
one of the example prompts is literally:
Prompt: in the style of a studio ghibli anime, a boy and his dog run up a grassy scenic mountain with gorgeous clouds, overlooking a village in the distant background
Wow that is dark, after Ghiblis staunch stance on AI.
These companies and their shareholders really are complete scum in my eyes, just like AI in miltech.
Not because the tech isn't super interesting but because they steal years of hard work and pain from actual artists with zero compensation - and then they brag about it in the most horrible way possible, with zero empathy.
Then comes losing the little humanity left the mainstream culture, exactly as Miyzaki said, leading to a dead cold and even more unjust society.
The Miyzaki quote is out of context, he isn't talking about generative AI but rather a 2016 animation of a creepy zombie whose limbs are controlled by AI.
While this is true, it's hard to imagine people spending years perfecting the style would be happy to see it copied effortlessly without any compensation while people who made the copying possible are rolling in cash.
This is not just about copyright infringement or plagiarism.
Automatically generating text, images and videos based on training data and a tiny prompt is fundamentally about taking someone's work and making money off of it without giving anything in return.
Kids are happy that homework takes less time. Teachers are happy that grading the generated homework takes less time. Programmers are happy they can write the same amount of code in less time. Graphic designers are happy they can get an SVG from a vague description immediately. Writers are happy they can generate filler from a few bullet points quickly.
But then someone comes along, notices people are not working most of the time, fires three quarters of them and demands 4x increased output from the rest. And they can do it because the "AI" is helping them.
Except they don't get paid any more. The company makes the same amount of money for less cost.
So where does the difference go? To the already rich who own the company and the product.
Power creates more power, money creates more money.
Communism is tossing the frog into boiling water (tens millions of dead), capitalism is boiling it slowly (poor people in first world countries might not afford a dentist but they're not starving yet).
We need a system that rewards work - human time and competence.
There are really only 2 resources in the world - natural resources and human time. Everything else is built on top of those. And the people providing their time should be rewarded, not those who are in positions of power which allow them to extract value while not providing anything in return.
56 minutes, 4 downvotes, HN is truly full of temporarily embarrassed millionaires.
Does anybody here really think rich people deserve to just get richer faster than any working person can? Does anybody really believe that buying up homes and companies and raking in money for doing absolutely nothing is what we should be rewarding?
Really impressive engineering work. The videos have gotten good enough that they can grab your attention and trigger a strong uncanny valley feeling.
I think OpenAI is actually doing a great job at easing people into these new technologies. It's not such a huge leap in capabilities that it's shocking, and it helps people acclimate for what's coming. This version is still limited but you can tell that in another generation or two it's going to break through some major capabilities threshold.
To give a comparison: in the LLM model space, the big capabilities threshold event for me came with the release of Gemini 2.5 Pro. The models before that were good in various ways, but that was the first model that felt truly magical.
From a creative perspective, it would be ideal if you could first generate a fixed set of assets, locations, and objects, which are then combined and used to bring multiple scenes to life while providing stronger continuity guarantees.
Impressively high level of continuity. The only errors I could really call out are:
1/ 0m23s: The moon polo players begin with the red coat rider putting on a pair of gloves, but they are not wearing gloves in the left-vs-right charge-down.
2/ 1m05s: The dragon flies up the coast with the cliffs on one side, but then the close-up has the direction of flight reversed. Also, the person speaking seemingly has their back to the direction of flight. (And a stripy instead of plain shirt and a harness that wasn’t visible before.)
3/ 1m45s: The ducks aren't taking the right hand corner into the straightaway. They are heading into the wall.
I do wonder what the workflow will be for fixing any more challenging continuity errors.
Not sure if it counts as a continuity error, but in the example "Prompt: Martial artist doing a bo-staff kata waist-deep in a koi pond", his wooden staff changes shape several times, resembling a bow at points. That was the first example I noticed as "clearly AI."
One use that occurred to me is that fans will be able to "fix" some movies that dropped the ball.
For example, I saw a lot of people criticizing "Wish" (2023, Disney) for being a good movie in the first half, and totally dropping the ball in the last half. I haven't seen it yet, but I'm wondering if fans will be able to evolve the source material in the future to get the best possible version of it.
Maybe we will even get a good closure for Lost (2004)!
(I'm ignoring copyright aspects, of course, because those are too boring :D)
Or just going to the Goofs section of a movie on IMDB, and fix the trivial issues (e.g. car had cracked window in earlier scene, and suddenly a normal window in another scene).
My issue is that the copyright aspect are what prevents me from using this as much as I otherwise would.
About 6 months ago I asked a few different AIs if they could translate a song for me as a learning experience, meaning not a simple translation, but more a word by word explanation of what each word meant, how it was conjugated, any more musical/lyrical only uses that aren't common outside of songs, and so on. I was consistently refused on copyright grounds, despite this seeming a fair use given the educational nature. If I pasted a line of the lyrics at a time, it would work initially, but eventually I would need to start a new chat because the AI determined I translated too much at once.
So in this one, if I wanted to ask it to create a video of the moment in Final Fantasy 6 when the bad guy wins, or a video of the main characters of Final Fantasy 7 and 8 having a sword duel, would it outright refuse for copyright reasons?
It sounds like it would block me, which makes me lose a bit of interest in the technology. I could try to get around it, but at what point might that lead to my account being flagged as a trouble maker trying to bypass 'safety' features. I'm hoping in a few years the copyright fights on AI dies down and we get more fair use allowance instead of the tighter limitations to try to prevent calls for tighter regulation.
Anyone with access able to confirm if you can start this with a still image and a prompt?
The recent Google Veo 3 paper "Video models are zero-shot learners and reasoners" made a fascinating argument for video generation models as multi-purpose computer vision tools in the same way that LLMs are multi-purpose NLP tools. https://video-zero-shot.github.io/
It includes a bunch of interesting prompting examples in the appendix, it would be interesting to see how those work against Sora 2.
That could totally power next generation of green-screen techs. Generative actors may not find favorable response in the audiences; but SFX, decor, extras, environments that react to actors' actions - amazing potential.
There's less here than you think. Video games have already been procedurally generating environment art for quite some time, and film/tv are already leveraging that with giant screens that use Unreal Engine to create the backgrounds.
AI could be helpful here, but it's not clear that it is required or an improvement.
I don't get the autotune argument. It's like saying we shouldn't be using electronic instruments because it's not real or we shouldn't use digital audio instruments because they're not real etc.
It's just a way to get different kind of sound. It won't make you good tracks.
At least in terms of realism, the image generation field is at the realism line now. Single frame generation with Wan 2.1 / 2.2 (and others) for example, will get you realism.
Sora 1 was also lauded as being incredibly good at physics based on the early cherry-picked examples. The phrase "world simulator" was thrown around a lot. That didn't last long once people finally got their hands on it though.
Kind of wondersome if they will start to combine LLM generation with actual world models/GPU engines. Imagine that your model generates the wireframes, the Engine generates the physics and then another model fills in the actual visuals, and gaps... So you have realistic physics and gaps are filled in... Will also help with image retention more, if objects moved behind each other.
The demo on their homepage shows really bad physics. There’s a lot of it, but that doesn’t mean it’s correct. The hair of Sam looks like a paper cutout in almost every shot.
> You can sign up in-app for a push notification when access opens for your account.
You need to be in the US/Canada and wait for this notification, and when you get an invite you can start using it in the app and on sora.com. And apparently you get 4 more invite codes that you can share with anyone, e.g. Android users:
> Android users will be able to access Sora 2 via http://sora.com once you have an invite code from someone who already has access
It's wild that I have a paid account but I have to scour the Internet to find someone else with a paid account and beg them for an invite code to use the product I already paid for. Make it make sense.
I know the comments here are gonna be negative but I just find this so sick and awesome. Feels like it's finally close to the potential we knew was possible a few years ago. Feels like a pixar moment when CG tech showed a new realm of what was possible with toy story
These videos are a very impressive engineering feat. There are a lot of uses for this capability that will be beneficial to society, and in the coming years people will come up with more good uses nobody today has thought of yet.
But clearly we also see some major downsides. We already have an epidemic of social media rotting people's minds, and everything about this capability is set to supercharge these trends. OpenAI addresses some of these concerns, but there's absolutely no reason to think that OpenAI will do anything other than what they perceive as whatever makes them the most money.
An analogy would be a company coming up with a way to synthesize and distribute infinite high-fructose corn syrup. There are positive aspects to cheaply making sweet tasting food, but we can also expect some very adverse effects on nutritional health. Sora looks like the equivalent for the mind.
There's an optimistic take on this fantastic new technology making the world a better place for all of us in the long run, after society and culture have adapted to it. It's going to be a bumpy ride before we get there.
I actually wonder if this will kill off the social apps and the bragging that happens. It will be flooded by people faking themselves doing the unimaginable.
Yes, I wonder if the content distribution networks that call themselves "social networks" can even survive something like this.
Of course, the ones focusing on the content can always editorialize the spam out. And in real social networks you ask your friends to stop making that much slop. But this can be finally the end of Facebook-like stuff.
This is also my thesis. The internet is going to be saturated with AI slop indiscernible from real content. Once it reaches a tipping point, there will no longer be much of a reason to consume the content at all. I think social networks that can authenticate video/photo/text content as human-created will be a major trend in a few years.
I have no clue if the reactions are real, but there are some videos online of people showing their grandparents gameplay from Grand Theft Auto games trying to convince them that it is real footage. The point of the videos is to laugh at their reactions where they question if it really happened, etc.
Maybe this will result in something similar, but it can affect more people who aren’t as wary.
I regularly get AI movie recaps on my shorts and I just eat it up.
The very fact that I (or billions of others) waste time on shorts is an issue. I don't even play games anymore, it's just shorts. That is a concerning rewiring of the brain :/
Guess what I`m trying to say is that, there is a market out there. It's not pretty, but there certainly is.
Will keep trying to not watch these damn shorts...
> There are a lot of uses for this capability that will be beneficial to society
Please enlighten me. What are they? If my elderly grandma is on her deathbed and I have no way to get to see her before she passes, will she get more warmth and fond memories of me with a clip of my figure riding an AI generated dragon saying goodbye, or a handwritten letter?
I still feel this is limited by what it learned from. It looks cool but it also looks like something I'd dreamt or saw flicking through TV channels. Kind of like spam for the eyes.
No doubt they can create Hollywood quality clips if the tools are good enough to keep objects consistent, example, coming back to the same scene with same decor and also emotional consistency in actors
I think this is not nearly as important as most people think it is.
In hollywood movies, everyone already knows about "continuity errors" - like when the water level of a glass goes up over time due to shots being spliced together. Sometimes shots with continuity errors are explicitly chosen by the editor because it had the most emotional resonance for the scene.
These types of things rarely affect our human subjective enjoyment of a video.
In terms of physics errors - current human CGI has physics errors. People just accept it and move on.
We know that superman can't lift an airplane because all of that weight on a single point of the fuselage doesn't hold, but like whatever.
Location consistency is important. Even something as simple and subtle as breaking the 180-rule [1] feels super uncanny to most audiences. Let alone changing the set the actor occupies, their wardrobe, props, etc.
There are lots of tools being built to address this, but they're still immature.
Well put. Honestly the actor part is mostly solved by now, the tricky part is depicting any kind of believable, persistent space across different shots. Based off of amateur outputs from places like https://www.reddit.com/r/aivideo/, at least!
This release is clearly capable of generating mind-blowingly realistic short clips, but I don't see any evidence that longer, multi-shot videos can be automated yet. With a professional's time and existing editing techniques, however...
I wonder if this stuff is trained on enough Hallmark movies that even AI actors will buy a hot coffee at a cafe and then proceed to flail the empty cup around like the humans do. Really takes me out of the scene every time - they can't even put water in the cup!?
No way man, this is why i loved Mr Robot, they actually payed a real expert and worked story around realism and not just made up gobbleygook that shuts my brain off entirely to its nonsense
The ability for the masses to create any video just by typing, among the other features, is not novel technology? Or is it just the lack of emotional response?
It's the inaccuracy of things like shadows, sub-surface scattering and specular highlights. I think the shadow inaccuracy is what the human visual system is most sensitive to.
These LLMs might make content that looks initially impressive but they are absolutely not performing physically based rendering or have any awareness of the lighting arrangement in these scenes. There are a lot of things they get right, but you only have to screw up one small element to throw the whole thing off.
I am willing to bet that Unreal Engine 5 will continue to produce more realistic human faces than OAI ever can with these types of models. You cannot beat the effects of actually running raytracing in a PBR pipeline.
Yeah, the faces aren't right, and impressive as it is I'm getting icky "uncanny valley" vibes from this.
CGI for fantasy stuff is unavoidable, but when it's stuff that could have been done by actors but is instead AI, then to me it just feels cheap and nasty - fake.
Sam Altman has made (for me) encouraging statements in the past about short-form video like TikTok being the best current example of misaligned AI. While this release references policies to combat "Doomscrolling and RL-sloptimization", it's curious that OpenAI would devote resources to building a social app based on AI generated short form video, which seems to be a core problem in our world. IMO you can't tweak the TikTok/YouTube shorts format and make it a societal good all of a sudden, especially with exclusively AI content. This is a disturbing development for Altman's leadership, and sort of explains what happened in 2023 when they tried to remove him... -> says one thing, does the opposite.
I'm optimistic about the Sora app! My hope is that it becomes much more whimsical and fun than TikTok because everyone on the app knows that all content is fake. Hopefully that means less rage-bait and more creative content, like OG YouTube. Nobody's going to get their news from Sora because it's literally 100% fake.
In the early years everyone told me that TikTok is actually fun and whimsical (like just after it stopped being musical.ly), and it's all about fun collaboration, and amateur comedy sketches, fun dances and lipsyncs, and people posting fun reactions to each other etc, all lighthearted and that social media is finally fun again!
Why would it be more like OG YouTube, when the content they demoed very closely resembles YouTube shorts? The key difference is OG YouTube was long form.
>IMO you can't tweak the TikTok/YouTube shorts format and make it a societal good all of a sudden, especially with exclusively AI content.
I agree. At best, short videos can be entertainment that destroys your attention span. Anything more is impossible. Even if there were no bad actors producing the content, you can't condense valuable information into this format.
Sam Altman is a businessman. His job is to say whatever assuages his market, and that includes gaslighting you when you're disgusted by AI.
If you never expected Altman to be the figurehead of principled philosophy, none of this should surprise you. Of course the startup alumni guy is going to project maligned expectations in the hopes of being a multi-trillion dollar company. The shareholders love that shit, Altman is applying the same lessons he learned at Worldcoin to a more successful business.
There was never any question why Altman was removed, in my mind. OpenAI outgrew it's need for grifters, but the grifter hadn't yet outgrown his need for OpenAI.
OpenAI apparently assumes that the primary users of Sora 2/the Sora app will be Gen Z, especially with the demo examples shown in the livestream. If they are trying to pull users from TikTok with this, it won't work: there's some nuance to Gen Z interests than being quirky and random, and if they did indeed pull users from TikTok then ByteDance could easily include their own image/video generators.
Sora 2 itself as a video model doesn't seem better than Veo 3/Kling 2.5/Wan 2.2, and the primary touted feature of having a consistent character can be sufficiently emulated in those models with an input image.
Good point. I think OpenAI lacks the cultural understanding that tiktok is providing their users not only with entertainment but also social things like trends, reviews, gossip, self-expression. These aspects are not included in the sora experience.
This is going to sound crass but idc - OAI is just full of geeks, when what is needed is people who are more akin to hippies - thats pretty much what Apple was in the early days.
Its no use building technology when its not married with the humanities and liberal arts.
I'm a software engineer and hobbyist actor/director. My friends are in the film industry and are in IATSE and SAG-AFTRA. I've made photons-on-glass films for decades, and I frequently film stuff with my friends for festivals.
I love this AI video technology.
Here are some of the films my friends and I have been making with AI. These are not "prompted", but instead use a lot of hand animation, rotoscoping, and human voice acting in addition to AI assistance:
I see several incredibly good things happening with this tech:
- More people being able to visually articulate themselves, including "lay" people who typically do not use editing software.
- Creative talent at the bottom rungs being able to reach high with their ambition and pitch grand ideas. With enough effort, they don't even need studio capital anymore. (Think about the tens of thousands of students that go to film school that never get to direct their dream film. That was a lot of us!)
- Smaller studios can start to compete with big studios. A ten person studio in France can now make a well-crafted animation that has more heart and soul than recent by-the-formula Pixar films. It's going to start looking like indie games. Silksong and Undertale and Stardew Valley, but for movies, shows, and shorts. Makoto Shinkai did this once by himself with "Voices of a Distant Star", but it hasn't been oft repeated. Now that is becoming possible.
You can't just "prompt" this stuff. It takes work. (Each of the shorts above took days of effort - something you probably wouldn't know unless you're in the trenches trying to use the tech!)
For people that know how to do a little VFX and editing, and that know the basic rules of storytelling, these tools are remarkable assets that compliment an existing skill set. But every shot, every location, every scene is still work. And you have to weave that all into a compelling story with good hooks and visuals. It's multi-layered and complex. Not unlike code.
And another code analogy: think of these models like Claude Code for the creative. An exoskeleton, but not the core driving engineer or vision that draws it all together. You can't prompt a code base, and similarly, you can't prompt a movie. At least not anytime soon.
Creative people with ambition and limited resources make good things today without this technology. All this does is accelerate the rate at which low quality "content" is produced by people that have no interest in learning a craft, without attribution and without compensation for the people that have made the effort and whose works train these models.
"The studios and creators who thrive in this new landscape will be those who can effectively harness AI’s capabilities while maintaining the human creativity and vision that ultimately drives the art of cinema."
It is in many ways thrilling to see this come to life, and I couldn't agree with you more.
> "The studios and creators who thrive in this new landscape will be those who can effectively harness AI’s capabilities while maintaining the human creativity and vision that ultimately drives the art of cinema."
..Just somehow several years on, these optimistic statements still all end up being in the future tense, somehow for all the supposed greatness and benefits, we still dont see really valuable outputs.
A lot of us do not want more of the "CONTENT" as envisioned by corporate ghouls who want their employees or artists to "thrive" (another word kidnapped by LinkedIn-Linguists). The point is not in the speed and easiness of generation of outputs, visual and sound effects etc. The point is the artists interpretation and their own vision, impressions etc. Not a statistical slop which "likely" fits my preferences (i.e. increases my dopamin levels).
Adding to the list: The Adventures of Reemo Green. Very funny, and the first time I’ve watched AI video and enjoyed it as more than a technical curiosity.
Carter needs a new laptop. His daily driver has been falling apart for ages but he refuses to give it up.
We all told him about the sound mix - he let a couple of videos slip with a bad "mono as single-channel stereo audio" renders. On his machine it sounded normal. He got flack for that, and he's been hearing this for months.
I'm going to show him this thread. I don't think he'll ever forget to check again.
Despite that, he's a really talented guy. Chalk this up as a bad production deploy. We didn't want to delete and re-upload since the videos had legs when we first released them. There's a checklist now.
I'll have you know that in this year's Atlanta 48 Hour Film project (something I've been doing since I was a teen), several teams used AI.
Rewind to just one year prior -- 2024.
AI video was brand-spanking new. We'd only just gotten over the "Will Smith" spaghetti video and the meme-y "Pepperoni Hug Spot" and "Harry Potter by Balenciaga" videos.
I was the only person to attempt to use AI in 2024's competition. It was a time when the tools and infrastructure for video barely existed.
On the debut night, I was resoundingly booed by the audience. It felt surreal. Working all weekend to have an audience of peers jeering at you in a dark theater. The judges gave me an award out of sympathy.
Back then, image-to-video models really were not a thing (Luma launched "Dream Machine v1" shortly after this). I was using Comfy, Blender, Mocap, a full Mocap suit (the itchy kind), and a lot of other hacks to build something with extremely crude tools.
We lost a day of filming and had to scramble to get something done in just 24 hours. No sleep, too much caffeine. Lots of sweat and toil.
The resulting film was a total mess, of course:
https://vimeo.com/955680517/05d9fb0c4f (It's seriously bad - I hate it. It might legitimately be the very first time AI was used in a 48 hour competition.)
That said, it felt very much like a real 48 Hour competition to me. Like a game jam. The crude ingredients, an idea, the clock. The hustle. The corners being cut. It was palpable.
I don't think you can say there isn't soul in this process. The process has so much soul.
Anyway, fast forward to this year. Three teams used AI, including my own. (I don't think I have a link to our film, sadly.)
We all got applause. The audience was full of industry folks, students, and hobbyists. They loved it. And they knew we used AI.
The industry is anxious but curious about the tech. But fundamentally, it's a new tool for the tool box. The real task is storytelling.
Going to be an amazing source of training data, wait till they get it to real time and people are leaving their video camera open for AR features. OpenAI is about to have a lot of current real world image data, never mind the sentiment analysis.
I don't think they were limited for video training data. Gathering real world data is pretty easy, gathering curated information is a little more difficult.
> A lot of problems with other apps stem from the monetization model incentivizing decisions that are at odds with user wellbeing. Transparently, our only current plan is to eventually give users the option to pay some amount to generate an extra video if there’s too much demand relative to available compute. As the app evolves, we will openly communicate any changes in our approach here, while continuing to keep user wellbeing as our main goal.
> And we're introducing Cameo, giving you the power to step into any world or scene, and letting your friends cast you in theirs.
How much are they (and providers of similar tools) going to be able to keep anyone from putting anyone else in a video, shown doing and saying whatever the tool user wants?
Will some only protect politicians and celebrities? Will the less-famous/less-powerful of us be harassed, defamed, exploited, scammed, etc.?
it seems like this is basically youtube's ContentID, but for your face. as long as you upload your "cameo" aka facial scan to them, they can recognize and control the generation of videos with it. if you don't give them your face, then they can't/won't.
"Consent-based likeness. Our goal is to place you in control of your likeness end-to-end with Sora. We have guardrails intended to ensure that your audio and image likeness are used with your consent, via cameos. Only you decide who can use your cameo, and you can revoke access at any time. We also take measures to block depictions of public figures (except those using the cameos feature, of course). Videos that include your cameo—including drafts created by other users—are always visible to you. This lets you easily review and delete (and, if needed, report) any videos featuring your cameo. We also apply extra safety guardrails to any video with a cameo, and you can even set preferences for how your cameo behaves—for example, requesting that it always wears a fedora."
Until you have 2 people that are near identical. They don’t even have to be twins, there are plenty of examples where people can’t even tell other people apart. How is an AI going to do it?
You don’t own your likeness. It’s not intellectual property. It’s a constantly changing representation of a biological being. It can’t even be absolutely defined— it’s always subject to the way in which it was captured. Does a person own their likeness for all time? Or only their current likeness? What about more abstract representations of their likeness?
The can of worms OpenAI is opening by going down this path is wild. We’re not current able to solve such a complex issue. We can’t even distinguish robots from humans on the internet.
If this company's guardrails end up sufficiently working well in practice (note phrases like "intended", "take measures", and "preferences...requested", on things they can't do 100%)... there will be weak links elsewhere, letting similar computation be performed without sufficiently effective guardrails against abuse?
How do we prepare for this? Societal adjustment only (e.g., disbelieving defamatory video, accepting what pervs will do)? Establishing a common base of cultural expectations for conduct? Increasing deterrence for abusers?
If Deepfakes remain the tools of nation state actors, laypeople will be easily fooled.
If Deepfakes are available on your iPhone and within TikTok, everyone will just ask "Is it Photoshop?" for every shred of doubt. (In fact, I already see people saying, "This looks like AI".)
This is good. Normalize the magic until it isn't magic anymore.
People will get it. They're smart. They just need exposure.
> People will get it. They're smart. They just need exposure.
I really doubt this.
If you are in the creative field, your work will just be reduced to "is this slop?" or "fixed it!" with a low effort AI generated work of your original work (fuck copyright right?).
I already see artists battling and fighting putting out their best non AI work only for their audience to question if it is real and they lose the impressiveness.
This just already undermines creators who don't use AI generated stuff.
But who cares about them right? "it is the future" and it is most definitely AGI for them.
But then again, the starving artist never really made any money and this ensures that the artform stays dead.
While the quality of what I'm seeing is very nice for AI generated content (I still can't believe it) but the fact thay they are mostly showing short clips and not a long connected consistent video makes it less impressive.
The result of that isn't rational skepticism, though. It's distrusting mainstream news and embracing whatever conspiracy theories your friends believe.
State of the things with doom scrolling was already bad, add to it layoffs and replacing people with AI (just admit it, interns are struggling competing with Claude Code, Cursor and Codex)
What's coming next? Bunch of people, with lots of free time watching non-sense AI generated content?
I am genuinely curious, because I was and still excited about AI, until I saw how doom scrolling is getting worse
>What's coming next? Bunch of people, with lots of free time watching non-sense AI generated content?
Wasn't this always the outcome of the post labor economy?
For this discussion lets just say that AI+Robots could replace most human labor and thinking. What do people do? Entertainment is going to be the number one time consumer.
I’m wondering how they really prevent uploads of other peoples faces if they take a clip of a video of another person. I’m sure Apple didn’t open up the 3d Face ID scanning to them to verify
IDK if the site is being hugged to death but I can only load the first video. Even in just one viewing there were noticeable artifacts, so my impression is that Veo is still in the lead here.
A little tangential to this announcement, but is anyone aware of any clean/ethical models for AI video or image generation (i.e. not trained on copyright work?) that are available publicly?
VFX artist and developer here, who's deep into this stuff, and it is really not there. It's an island of itself, barely controllable and barely usable with other media. They are just now getting around to generating alpha channels, with virtual none of the existing pipelines for any AI video or image generation tools to even incorporate and work with alpha channels. This is just one of several hundred aspects of incompatibility. It really seriously appears as of no one at any of the AI video generation research teams has any professional media production experience, or even bothered too look at existing media production data standards, and what they are making tool-wise is incompatible in every possible respect.
It's insanely impressive. At the same time, all these videos all look terrible to me. Still get extreme uncanny valley and literally makes me sick to my stomach.
This stuff works really well when you make something that's exaggerated reality, as in either an animation or a MTV-style music video
I can't find the link now, but I saw a continuous shot video of a grocery store from the perspective of a fly. It was shot in the 90s music video style and looked so damn good.
Some of the stuff being done by these guys is also a whole lot of fun (slightly NSFW and political content), and it fits the music video theme:
Agree - leaps and bounds beyond anything I would have dreamed possible a few years ago...but... IDK, if I'm honest, the sound was way off too, not just the visuals. The music sounded detuned slightly, and the crowd noise was "crackly" etc. etc. It had a low-fidelity "quality" to it.
Personally, I feel mixed feelings. I'm impressed, but I'm not looking forward to the new "movies" that are going to litter YouTube et al generated from this.
I love the casual reminds that we're second-class citizens each time a new technology gets released. Available in the US but always excluding Puerto Rico.
I saw a famous actor-director (can't remember who, but an A-list guy) said it would be super valuable even if you only use it for establishing shots.
Like you have an exterior shot of a cabin, the surrounding environment, etc — all generated. Then you jump inside which can be shot on a traditional set in a studio.
Getting that establishing shot in real life might cost $30K to find a location, get the crew there, etc. Huge boon to indie films on a budget, but being able to endlessly tweak the shot is valuable even for productions that could afford to do it IRL.
Searched around and found it. It was actually Ashton Kutcher's interview with Eric Schmidt.
Kutcher mentions the establishing shots, and I'd forgotten also points out the utility for relatively short stunt sequences.
> Why would you go out and shoot an establishing shot of a house in a television show when you could just create the establishing shot for $100? To go out and shoot it would cost you thousands of dollars.
> Action scenes of me jumping off of this building, you don’t have to have a stunt person go do it, you could just go do it [with AI].
Wow. What an intelligent take. I would have never expected this from Ben Affleck. He seems extremely familiar with the technology and it's capabilities and limits.
Casey Affleck is currently shooting a horror vampire period piece using Comfy UI and an Unreal Engine Volume. The AI is used for the background plates. It's just a test, but it's happening right now.
People use these for sure but the biggest problem with these I feel is that they produce ”finished shots” with 8 bit colors and heavy grading. It’s hard to mix it with the other material which actually looks quite bland while it is being worked on. Would be great if somebody would train a model on raw footage.
I can see it being interesting to create wacky fake videos of your friends for a week or two, but why would people still be using this next year?
I watch videos for two reasons. To see real things, or to consume interesting stories. These videos are not real, and the storytelling is still very limited.
Does this survive panning the camera away for 5 to 10 seconds and then back? Or basic conversation scene with the camera cutting between being located behind either speaker once every few seconds?
Basically proper working persistence of the scene.
Dude, this generation of AI video models are just starting to have basic camera production terms understood, and then it is exactly like LLM generation: it's a pull of a slot machine arm; you might get what you want, but that's "winning" and the slot machine only gives out winners one in every 100 pulls. Every possible thing that could not be right happens.
For example, I'm working with a walking and talking character at this time using multiple AI video models and systems. Generated clips any length longer than 8 seconds risk rapid quality loss, but sometimes you can get up to 12-19 seconds without the generation breaking down. That means one needs to simulate a multiple camera shoot on a stage, so you can cut around the character(s) and create a longer sequence. But now you need to have multiple views of the same location to place your character(s) into - and current AI models can't reliably give you a "different angled views" of an environment. We just got consistent different views of characters, and it'll be another period until environments can be generally examined from any view. BUT, that's if people realize this is not in the models yet, and so far people are so fascinated by the fantasy violence and sexual content they can make nobody realizes you cannot simply "look left and right" in any of these models and that even works with consistency or reliability. There are workarounds, like creating one's entire set and environments in 3D models, for use as the backgrounds and starting frames, but that's now 3D media production + AI, and none of the AI tools generate media that even has alpha channels, and a lot of similar incompatibilities like that.
This is super cool and fun and will almost certainly be really bad for society in loads of different ways. From the descriptions of all the guardrails they're needing to put in it seems like they know it too.
Glad to see someone is looking out for a forest, here. A diverse host of excuses have cropped up to explain away the anxiety AGI brings, and I totally understand why. Yet again, today we stare into the abyss.
Sora 2 represents significant progress towards [AGI]. In keeping with OpenAI’s mission, it is important that humanity benefits from these models as they are developed.
I wonder how exactly they reconcile the quote above with "We are concerned about late-stage AGI development becoming a competitive race without time for adequate safety precautions"...
I am not for or against AGI, but why is there anxiety around it? Do people simply hear sales rhetoric and assume that it can exist and will be used in order to dominate their lives?
So being able to generate sign language videos for people who cannot hear is not a legitimate use case for AI videos? Or is your hate boner for AI just blinding you from useful applications?
Great point. Really, the main problem with subtitles is that the creator can understand them without having to know another language, and therefore can spot check them. That makes it much more difficult to insert Black Mirror-style Contextually Relevant Advertisements.
sign languages are completely different languages from spoken languages, with their own grammar etc.
subtitles can work but it's basically a second language. perhaps comparable to many countries where people speak a dialect that's very different from the "standard" written language.
this is why you sometimes have sign language interpreters at events, rather than just captions.
there's not really a widely accepted written form of sign language.
>this is why you sometimes have sign language interpreters at events, rather than just captions.
No, the reason is because a) it's in real time, and b) there's no screen to put the subtitles on. If it was possible to simply display subtitles on people's vision, that would be much more preferable, because writing is a form of communication more people are familiar with than sign language. For example, someone might not be deaf, but might still not be able to hear the audio, so a sign language interpreter would not help them at all, while closed captions would.
if you're maximizing accessibility you'd have both. often in broadcasts with closed captioning, there will also be a video of the sign language interpreter.
Shameless plug but I am creating a startup in this space called cleanvideo.cc to tackle some of the issues that will come with fake news videos. https://cleanvideo.cc
I really hope they have more granular APIs around this.
One use case I'm really excited about is simply making animated sprites and rotational transformations of artwork using these videogen models, but unlike with local open models, they never seem to expose things like depth estimation output heads, aspect ratio alteration, or other things that would actually make these useful tools beyond shortform content generation.
in a computer graphics course i took, we looked through how popular film stories were tied to the technical achievements of that era. for example, toy story was an story born from the new found ability to render plastics effectively. similarly, the sora video seems to showcase a particular set of slow moving scenes (or when fast, disappearing into fluid water and clouds) which seem characteristic of this technology at the current moment in time
If I was on the OpenAI marketing team I maybe wouldn't have included the phrase "and letting your friends cast you in their [videos]". It's a little chilling.
What is the target market for this? The videos are not good enough for YouTube. They are unrealistic, nauseating and dorky. Already now any YouTube video that contains a hint of "AI" attracts hundreds of scathing comments. People do not want this.
Let me guess, the ultimate market will be teenagers "creating" a Skibidi Toilet and cheap TikTok propaganda videos which promote Gazan ocean front properties.
It's very hard for simultaneous good audio generation with video generation (simultaneous generation is necessary to maintain lip sync). Veo 3 et al also have flat monochannel audio, but not as bad as these Sora 2 demos.
Art should require effort. And by that I mean effort on the part of the artist. Not environmental damage. I am SO tired of non tech friends SWOONING me with some song they made in 0.3 seconds. I tell them, sarcastically, that I am indeed very impressed with their endeavors.
I know many people will disagree with me here, but I would be heart broken if it turned out someone like Nick Cave was AI generated.
And of course this goes into a philosophical debate. What does it matter if it was generated by AI?
And that's where we are heading. But for me I feel effort is required, where we are going means close to 0 effort required. Someone here said that just raises the bar for good movies. I say that mostly means we will get 1 billion movies. Most are "free" to produce and displaces the 0.0001% human made/good stuff. I dunno. Whoever had the PR machine on point got the blockbuster. Not weird, since the studio tried 300 000 000 of them at the same time.
Who the fuck wants that?
I feel like that ship in Wall-E. Let's invest in slurpies.
Anyway; AI is here and all of that, we are all embracing it. Will be interesting to see how all this ends once the fallout lands.
Sorry for a comment that feels all over the place; on the tram :)
Find this sort of innovation far less interesting or exciting than the text & speech work, but it seems to be a primary driver of adoption for the median person in a way that text capability simply is not.
However, personalization (teleporting yourself into a video scene) is boring to me. At its core, it doesn't generate new experience to me. My experience is not defined by photos / videos I took on a trip.
I also can't think of a reason why I would ever want to look at an AI generated video.
however as they hint at a little in the announcement, if video generation becomes good enough at simulating physics and environments realistically, that's very interesting for robotics.
It's obvious there is no way OpenAI can keep videos generated by this within their ecosystem. Everything will be fake, nothing real. We are going to have to change the way we interact with video. While it's obviously possible to fake videos today, it takes work by the creator and takes skill. Now it will take no skill so the obvious consequence of this is we can't believe anything we see.
The worst part is we are already seeing bad actors saying 'I didn't say that' or 'I didn't do that, it was a deep fake'. Now you will be able to say anything in real life and use AI for plausible deniability.
It's not that obvious. iOS is pretty secure, if they keep the social network and cameo feature limited to that there might not be good ways to export videos off the platform onto others beyond pointing a camera at the tablet screen. And beyond there being lots of ways to watermark stuff to be detectable, nothing stops the device using its own camera to try and spot if it's being recorded. The bar can be raised quite high as long as you're willing to exclude any device that isn't an iPhone/iPad.
Today's Sora can produce something that resembles reality from a distance, but if you look closely, especially if there's another perspective or the scene is atypical, the flaws are obvious.
Perhaps tomorrow's Sora will overcome the the "final 10%" and maintain undetectable consistency of objects in 2 perspectives. But that would require a spatial awareness and consistency that models still have a lot of trouble with.
It's also possible we remain stuck in the uncanny valley forever, or at least for the rest of our lives.
It's possible to produce some video or image that looks real, cherry-picked for a demo, but not possible to produce any arbitrary one you want that will end up passable.
We are going to have to change the way we interact with video.
I doubt it will be for the better. The ubiquity of AI deepfakes just reenforces entrenchment around "If the message reinforces my preconceived notion, I believe it and think anyone who calls it fake is stupid/my enemy/pushing an agenda. If the message contradicts my preconceived notion, it's obviously fake and anyone who believes it is stupid/my enemy/pushing an agenda.". People don't even take the time to think "is this even plausible", much less do the intellectual work to verify.
>Everything will be fake, nothing real. We are going to have to change the way we interact with video.
I'm optimistic here.
Look at 1900s tech like social security number/card, and paper birth certificates. Our world is changing and new systems of verification will be needed.
I see this as either terribly dystopian - or - a possibility for the mass expansion of cryptography and encrypted/signed communication. Ideally in privacy preserving ways because nothing else will make as much sense when it comes to the verification that countries will need to give each other even if they want backdoor registry BS for the common man.
Doing this as a social app somehow feels really gross, and I can't quite put to words why.
Like, it should be preferable to keep all the slop in the same trough. But it's like they can't come up with even one legitimate use case, and so the best product they can build around the technology is to try to create an addictive loop of consuming nothing but auto-generated "empty-calories" content.
I see it more as recognizing that it will take time to be good enough for other use cases so for this release they are targeting it as just something to have fun with. After seeing LLMs crammed into everything whether it makes sense or not, I can appreciate that.
It's obvious there is no way OpenAI can keep videos generated by this within their ecosystem. Everything will be fake, nothing real. We are going to have to change the way we interact with video. While it's obviously possible to fake videos today, it takes work by the creator and takes skill. Now it will take no skill so the obvious consequence of this is we can't believe anything we see.
The worst part is we are already seeing bad actors saying 'I didn't say that' or 'I didn't do that, it was a deep fake'. Now you will be able to say anything in real life and use AI for plausible deniability.
I predict a re-resurgence in life performances. Live music and live theater. People are going to get tired of video content when everything is fake.
I wonder how this will affect the large cinema production companies (Disney, WB, Universal, Sony, Paramount, 20th century...). The global film market share was estimated to be 100B in 2023. If the production cost of high FX movies like Avengers Infinity War goes down from 300M$ to just 10K$ in a couple of years, will companies like Disney restrain themselves to just release a few epic movies per year? Or will we be flooded with tons of slop? If this kind of AI content keeps getting better, how will movies sustain our attention and feel 'special'? Will people not care if an actor is AI or real?
This is the kind of thing people get excited about for the first couple of months and then barely use it going forward. It's amazing how quickly the novelty of this amazing technology wears off. You realize how necessary meaning/identity/narrative is to media and how empty it gets (regardless of the output) when those elements are missing.
Ok that's technically really impressive, and probably totally unusable in a real creativity context beyond stupid ads and politically-motivated deepfakes.
So, this is the AI Slop generator for the AI SlipSlop that Altman has announced lately.
Brave new internet, where humans are not needed for any "social" media anymore, AI will generate slop for bots without any human interaction in an endless cycle.
Cool - now let's see how much it costs in compute to generate a single clip. (Also, notice how no individual scene is longer than a handful of seconds?)
So far the true progress it has made is getting textures right close up. It still fudges how skin looks like the more it pans away from the characters.
Just seeing the examples that I assumed are cherry picked, it seems like they're still behind on Google when it comes to video generation, the physics and stylized versions of these shots seem not great. Veo3 was such a huge leap and is still ahead of many of the other large AI labs.
You can use the "cameo" feature only with users who have gone through the cameo creation flow. Sama has an account and created a cameo likeness of himself.
When you create your cameo you can choose who is allowed to make videos using it: "only me", "people I approve", "mutuals", or "everyone".
Someone who doesn't follow the moving edge would be forgiven for being confused by the dismissive criticism dominating this thread so far.
It's not that I disagree with the criticism; it's rather that when you live on the moving edge it's easy to lose track of the fact that things like this are miraculous and I know not a single person who thought we would get results "even" like this, this quickly.
This is a forum frequented by people making a living on the edge—get it. But still, remember to enjoy a little that you are living in a time of miracles. I hope we have leave to enjoy that.
I think HN is too political like this tech is clearly amazing and it’s great they shipped it there should be more props even if it’s a billion dollar company.
Yes the tech is amazing. But tech is not everything, after 20 years of social media, its pretty clear to everyone that those things can have large long term impact both positive and negative for society, discussing the potential impacts of the tech is not being "political", its just being interested in the future.
“With Sora 2, we are jumping straight to what we think may be the GPT‑3.5 moment for video.”
I think feeling like you need to use that in marketing copy is a pretty good clue in itself both that its not, and that you don’t believe it is so much as desperately wish it would be.
I am looking at the videos and really had a feeling that it looks right (minus a lot of obvious fuck ups still) where previously something felt fundamentally wrong with ai videos. It feels somewhat important, in so far you consider ai generated videos important.
It's the skin textures. It's the slightly better lipsyncing. Maybe it will be different when us normal users get it but so far the demos with Sam don't make him look waxy.
The Sora app squaring off against Meta's social video app is the real story here.
Sora 2 itself looks and sounds a little poorer than Google Veo 3. (Which is itself not currently ranked as the top video model. The Chinese models are dominating.)
I think Google, with their massive YouTube data set, is ultimately going to win this game. They have all the data and infrastructure in the world to build best-in-class video models, and they're just getting started.
The social battle will be something completely different, though. And that's something that I think OpenAI stands a good chance at winning.
Edit: Most companies that are confident of their image or video models stealthily launch it on the Model Arena a week ahead of the public model release. OpenAI did not arrange to do that for Sora 2.
Nano Banana, Seedream/Seedance, Kling, and several other models have followed this pattern of "stealth ELO ranking, then reveal pole position".
The fact that this model is about "friends" and "social" implies that this is an underpowered model. You probably saw a cherry picked highlight reel with a large VRAM context, but the actual consumer product will be engineered for efficiency. Built to sustain a high volume of cheap generations, not expensive high quality ones. A product built to face off against Meta. That model compete on the basis of putting you into videos with Pikachu, Mario, and Goku.
> I think Google, with their massive YouTube data set, is ultimately going to win this game.
I don't know, applying the same thinking to LLMs, Google should have been first and best with just text based LLMs too, considering the datasets they sit on (and researchers, among others the people who came up with attention). But OpenAI somehow beat them on that regardless.
I just had a thought: (spoilers Expanse and Hyperion and Fire Upon the Deep)
Multiple sci-fi-fantasy tales have been written about technology getting so out of control, either through its own doing or by abuse by a malevolent controller, that society must sever itself from that technology very intentionally and permanently.
I think the idea of AGI and transhumanism is that moment for society. I think it's hard to put the genie back in the bottle because multiple adversarial powers are racing to be more powerful than the rest, but maybe the best thing for society would be if every tensor chip disintegrated the moment they came into existence.
I don't see how society is better when everyone can run their own gooner simulation and share it with videos made of their high school classmates. Or how we'll benefit from being unable to trust any photo or video we see without trusting who sends it to you, and even then doubting its veracity. Not being able to hear your spouse's voice on the phone without checking the post-quantum digital signature of their transmission for authenticity.
Society is heading to a less stable, less certain moment than any point in its history, and it is happening within our lifetime.
As usual: impressive until you look close. Just freeze the frame and you see all the typical slop errors: pretty much any kind of writing is a garbled mess (look at the camera in the beginning). The horn of the unicorn sits on the bridle. The buttons on Sam's circus uniform hover in the air. There are candleholders with somehow candles inside as well as on top. The miniature instruments often make no sense. The conductor has 4 fingers on one hand and 5 on the other. The cheers of the audience is basically brown noise. Nedless to say, if you freeze the audience, hands are literally all over the place. Of course, everything conveniently has a ton of motion blur so you cannot see any detail.
I know, I know. Most people don't care. How exciting.
Is your complaint that it has errors? I mean look at what it can do. This is a freaking computer generating things from scratch based on a prompt. Two years ago, technology like this was so much worse and could only generate basic images and videos. Now it can generate visuals all from the text someone puts in.
Anyone, literally anyone, can use it (eventually) to generate incredible scenes. Imagine the person who comes up with a short film about an epic battle between griffins and aliens...Or a simple story of a boy walking in the woods with their dog...Or a story of a first kiss. Previously people were limited to what they had at hand. They couldn't produce a video because it was too costly. Now they can craft a video to meet their vision.
Well, yes? There's a reason why everything that was produced with these tools so far is garbage: because no one actually caring about their art would accept these things. Art is a deliberate thing, it takes effort. These tools are fine for company training videos and TikToks. Of course a few years ago this was science fiction. They are immensely impressive from a technical perspective. Two things can be true.
Now videos will be generated on the fly based on your preference. You will never put your phone down, it will detect when your sad or happy and generate videos accordingly
I don't know if it's just me or other people are feeling it as well. I don't enjoy videos anymore (unless live sports). I don't enjoy reading on my monitor anymore, I have been going back to physical books more often. I am in my early thirties.
The point is that sora2 demo videos seemed impressive but I just didn't feel any real excitement. I am not sure who this is really helping.
Personally I can't wait for super creative and novel indie film productions as film production will be more liberated from the grip of Hollywood and the influence of the upper classes in general. Especially once the Chinese make less-censored-to-Western-users models more available and even more so once people can run these things at home in some years.
Overall, appears rather underwhelming. Long way to go still for video generation.
Also, launching this as a social app seems like yet another desperate try to productize and monetize their tech, but this is the position big VC money forces you into.
I've perhaps been away from the scene for a bit, but I'm very impressed. To me this is absolutely "video generation", and I don't get your disdain for productization and monetization; last I checked this wasn't "Basic Research News".
I don't think it's disdainful to point out the lack of PMF for a dedicated app for Sora, nor how its behind competitors who don't require a dedicated social app. No need to strawman the guy, I think it's okay to be reasonably critical of ideas still on this website.
Inb4 make your own video model and see how easy it is
I guess copyright is pretty much dead now that the economy relies on violating it. Too bad those of us not invested into AI still won't be able to freely trade data as we please....
So a social network that's 100% your friends doing silly AI things?
I feel like this is the ultimate extension of "it feels like my feed is just the artificial version of what's happening my friends and doesn't really tell me anything about how they're actually faring."
Social media also tends to highlight the best parts of people’s lives, creating unrealistic expectations and views for those consuming it and looking at their real life. Now social media won’t even be a highlight reel, but completely fabricated.
I have to imagine there will be a rebellion against all of this at some point, when people simply can’t take the false realities anymore. What is the alternative? Ready Player One? The Matrix? Wall-E?
Maybe inside of a social network specially for AI, but a concerning number of people don't realize images and videos are AI, even when it's bad AI. As it gets better, and starts integrating the poster's image (like Sora 2), that's going to get even worse.
Some people use filters/photoshop to artificially juice their images; now they can use AI to artificially juice every aspect of their on-line presence.
I just watched the announcement video and something about it just gives me the ick. The whole time I just had the uncanny valley feeling.
The technology itself is super impressive, but a social media app of AI slop doesn't feel like the best use of it. I'm old enough to not really be interested in social media in general anymore, so maybe I'm just out of touch, but I just can't see this catching on. It feels like the type of thing that people will download, use a few times until the novelty wears off and then never open again.
I built an MVP of this [1] with images (not video) and in more of an Instagram style (not tiktok) back in ‘22, with the tagline ‘What if social media were literally fake?’
I am bullish on this, albeit with major concerns in many domains. It was fun and addictive as hell with images. With video it will be wild.
I find it comical that OpenAI with all the power of CharGPT even them are unable to release an app for both iOS and Android at the same time. Wow, good marketing for Codex.
Or Sama's documented reverence for Apple products. We are talking about the guy who sold Tim Cook his AI for $0.00, he's not exactly got the horse drawing the cart here.
Every AI video demonstration is always about funny stuff and fancy situations. We never see videos on art, history, literature, poetry, religion (imagine building a video about the moment Jesus was born) ... ducks in a race !? Come on ...
So much visual power, yet so little soul power. We are dying.
What do you imagine a generated video about poetry would be?
>Every AI video demonstration is always about funny stuff and fancy situations.
The thing about AI slop is that by its very nature, unless it's heavily reined in by a human, it's invariably lowest common denominator garbage. It very likely will generate something you yourself could think of within the first five seconds of hearing the prompt, not some very clever take on it, so it can only work as a placeholder (AI as a replacement of stock images is great, for example) or to add background detail where it won't call attention to itself and its genericity.
>imagine building a video about the moment Jesus was born
Given there are multiple paintings on the subject, I very much doubt no one has generated something like that already.
Before AI, each video on this channel would have taken a large team with a Hollywood budget to create. In a few more years, one person may be able to turn their creative vision into a full-length movie.
The value will shift to search or curation - if the cost to produce drops to nil, then the value will be in finding good content amongst a flood of sameness.
Things are cool because they are unique, very hard to create, and require creativity. When those things become cheap commodities, they are no longer cool.
The same could be said about software, and it's safe to say that open-source software making complex workflows easier and more efficient is a net good.
Making better tools is better for everyone: the median usage of those tools downstream is a separate issue.
Indeed. Art is partially evaluated by how impressive it is. That's why posting AI images on social media won't yield a lot of likes anymore. People have gotten used to images being easy to create, so they aren't seen as valuable anymore. The same will be true for videos.
AI pictures today are much less impressive than Dall-E 2 pictures were a few years ago, despite the fact that the models are much better nowadays. Currently AI videos can still be impressive, but this will quickly become a thing of the past.
Then people will move from trying to create art to creating "content". That is, non-artistic slop. Advertisements. Porn. Meme jokes. Click bait. Rage bait. Propaganda. Etc.
I would argue that we just get pickier and more sensitive to slop. When everyone can make a movie, the standard for a good movie will be higher. Many current Hollywood films wouldn’t make the cut. But maybe some kid in Nigeria makes the greatest film of all time.
No, but there's some stuff that are really creative. Ironically I think the reason I'm more positive about it is because I only encounter AI generated (non-text media) ~ once a week / 2 weeks.
Advertising: you (her) wearing new clothing before purchase, hair/glasses/makeup, make overs; guys after 3 months of gym membership, you driving the new car, you in this specific new home... etc, etc... I'm surprised this is not already everywhere, but people are too occupied making nsfw and fantasy violence clips.
Targeted advertising has become just manipulation. I don't know if personalized advertisement videos for everyone promoting a fake world that doesn't exist is really a benefit for the world...
If course it's not a benefit, but it's an advertising angle that will work very well with a class of gullible consumers, and that is enough to justify it being plastered everywhere. I don't write these rules, I just notice them.
Democracy? Strengthened! Nothing says “informed electorate” like not knowing if a politician actually said they support nazism or if it was just a hyper-realistic AI puppet.
Trust in media? Soaring! Why believe your eyes or ears when you can doubt everything equally?
Journalism? Thriving! Reporters now get to spend their days playing forensic video detective instead of, you know, reporting news.
Social harmony? Better than ever! Nothing brings people together like shared paranoia and the collective shrug of “I guess truth is dead now.”
lol i wonder if this will create a market for PKI at the image sensor level so that videos will be cryptographically signed and baked into the actual video stream with steganography.
I imagine it's incredibly useful for prototyping movies, tv, commercials before going to the final version. CGI will probably get way cheaper too with some hybrid approach.
Obviously this will get used for a lot of evil or bad as well
I feel like that's missing the point of pre-vis anyway, its purpose is to lay down key details with precision but without regard for fidelity (e.g. https://youtu.be/KMMeHPGV5VE), a system with high fidelity but very loose control is the exact opposite of what they want.
... how dare you, sir. That is entirely unacceptable and you will be reported to the ministry of proper living!
Regardless of the slop, some people will learn to use it well. You have stuff like NeuralViz - quite the sight! - and other creators will follow suit, and figure out how to use the new tools to produce content that's worth engaging with. Bigfoot vlogs and dinosaur chase scenes, all that stuff is mostly just fun.
People like to play. Let them play. This stuff looks fun, and beats Sora 1 by a long shot.
Could you please stop posting unsubstantive comments and flamebait? You've unfortunately been doing it repeatedly. It's not what this site is for, and destroys what it is for.