Resources:
Join the discussion: Got something to say? Let us know on LinkedIn and network with other AI leaders
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Try Our Free AI Prompting Course: Register for our free Prime, Prompt, and Polish AI Course!
Sora 2: The New AI Video Tool That Quietly Delivers a Business Advantage
OpenAI’s recent release of Sora 2 has drawn headlines for its realism, release strategy, and the ethical debates it fuels around social media and misinformation. Beneath the media noise, however, Sora 2 unlocks unique, actionable opportunities that distinctively benefit business leaders. Here is a closer look — with specifics — at how Sora 2 stands apart, its technical advances, and why its practical applications could provide a tangible edge for organizations focused on brand building, operational efficiency, and creative content.
From AI Hype to Practical Utility: What Sets Sora 2 Apart
Sora 2 is currently the leading AI video model, according to detailed analysis of its capabilities. Unlike its predecessor, Sora 2 generates hyper-realistic videos from text prompts with significant upgrades: physics-aware simulation, synchronized audio (including music and dialogue), and user-editable AI avatars called "Cameos." These features lead to highly realistic outputs that surpass even the most advanced competitors, such as Google’s VO3.
Sora 2 operates through an invite-only iOS app as well as a web portal. Remarkably, access is available without requiring a paid plan — the limiting factor is an invite code. While AI copyright and misinformation concerns are prominent in public discourse, these are not the areas where Sora 2 currently delivers its greatest business value.
Technical Advancements Drive Quality and Efficiency
The technical leap in Sora 2’s realism is not incremental. It achieves physics simulation in video outputs that were previously unattainable, even with repeated prompts on earlier models. Video generations now routinely display realistic human and object movement, accurate environmental interaction, and consistent lighting across multiple camera angles. This is more than a cosmetic upgrade: it reduces rework, shortens production cycles, and increases the reliability of AI as a creative partner.
The synchronized audio integration is similarly advanced. Sora 2 can align ambient sounds, dialogue, and music to the video, removing the need for cumbersome multi-tool workflows. This allows for the rapid creation of multi-sensory content, leaving only minimal polishing required — a critical benefit for marketing, training, and communications teams tasked with high-volume content needs.
AI Avatars and the Cameo Feature: A New Lever for Brand Storytelling
Sora 2 introduces a self-serve “Cameo” feature, allowing users to create hyper-realistic video avatars of themselves after a quick in-app verification. Only the individual has control over their avatar’s usage, and notifications are provided each time it appears in a video. This grants business leaders, executives, and subject matter experts the ability to scale their personal presence — or that of specific company representatives — across a wide range of content, from executive messages to specialized training modules or product demos.
With cameo videos, leadership teams can appear anywhere, in any scenario, and deliver consistent branded messages without the logistical hassle of travel, expensive video shoots, or the time investment of repeated recording sessions. Companies can create diverse, customized content at scale by leveraging Sora’s editing tools (e.g., Remix) and scene flexibility.
Social Media Virality vs. Brand Protection
While Sora 2 is built with a TikTok-style interface and includes features that reward content remixing and sharing, its business upside is found in private, controlled content creation. The system’s watermarking, both visible (animated, moving watermarks) and invisible (metadata via the Coalition for Content Provenance and Authenticity), presents some hurdles to unauthorized reproduction. Yet, simple editing techniques can often circumvent these safeguards, a fact professionals must consider for risk management and brand protection.
The platform allows only verified users to create their own avatars, and control mechanisms are in place to restrict or expand who can use each cameo. Businesses can tightly manage their executives’ digital likenesses, avoiding many of the deepfake risks present in open, unregulated AI video solutions.
Copyright, Compliance, and the Future of Digital Content
Sora 2’s approach to copyright is aggressive and may subject OpenAI to legal challenge — the company’s default is opt-out policies for training data, with individual infringement reporting required from copyright holders. The presence of widely recognizable copyrighted characters and IP in the Sora feed raises questions that legal and compliance teams will need to track as this technology matures.
However, for businesses producing original content and controlling their own IP, Sora 2’s value is less about liability than about seizing a rare window for creative differentiation. The risk-reward calculus shifts as the quality, speed, and versatility of output enables faster iteration and experimentation, with creative control remaining in the hands of the company.
Translating Sora 2’s Potential Into Business Results
The platform’s “cheat code” lies in combining the cameo feature with Sora’s web-based, multi-format video output. For firms willing to invest in CEO or leadership avatar creation, the possibilities include:
High-frequency, hyper-personalized executive communications
Localization of marketing campaigns without the need for new shoots
Effortless creation of explainer videos, training content, or even virtual tours
Scalable social media content by repurposing a single cameo in countless creative formats
Sora 2 removes many typical barriers: video shoots, travel, editing time, and even on-camera availability. Businesses can now ideate and deploy professional-grade video content, at speed and scale, controlled by internal teams.
Conclusion: An Unfair Advantage for the Early Movers
While much attention will remain focused on Sora 2’s social impact and the risks of viral AI slop or misinformation, the immediate, tangible business value is clear for those prepared to act. The technology grants companies a means to accelerate content pipelines, inject executive presence into campaigns or learning programs, and create brand consistency across digital channels — all while controlling cost and minimizing friction.
The prime opportunity is now, before industry-wide saturation and legal frameworks catch up. Sora 2 is not simply another viral trend: it is an accessible tool for any organization ready to experiment with the next frontier of business communications and brand-building.
Topics Covered in This Episode:
- Sora 2 AI Video Model Overview
- Sora 2's Hyper-Realistic AI Avatars
- Sora 2 Tech Improvements: Audio & Physics
- Sora 2 iOS App Release Strategy
- Sora 2's Cameo AI Avatar Feature
- Sora 2 Remix and Fork Video Tools
- Sora 2 Copyright and Deepfake Risks
- Sora 2 Animated & Invisible Watermark Protection
- Sora 2's Impact on Short-Form Content Creators
- Sora 2 Business Applications and Brand Building
- Sora 2 Physics-Aware Video Generation Demo
- Sora 2's Role in Doomscrolling and AI Slop
- Sora 2 Social Media Integration and Viral Potential
- Sora 2's Future in Business Marketing Strategies
Episode Keywords
Sora 2, OpenAI, AI video model, AI avatars, business utility AI, iOS social media app, invite only app, TikTok-style AI app, AI brain rot, disinformation, copyright issues AI, AI legal headache, deepfake risks, Cameo feature, video avatar, verified users, remix feature, generative AI video, physics aware AI simulation, synced audio, AI video realism, temporal consistency, world simulation AI, text to video, hyper-realistic videos, soundscaping AI, ambient sound generation, dialogue synchronization, Sora watermark, AI watermarking, invisible metadata watermark, CP2A metadata, Sora free tier, Sora pro plan, ChatGPT credentials, Sora API, user permissions AI avatar, AI content moderation, AI-generated disinformation, viral video app, TikTokification of AI, short-form content generation, autoplay algorithm AI, engagement patterns AI, doomscrolling, AI slop, TikTok competitor AI, YouTube Shorts AI, Reels AI, social media AI tools, AI marketing cheat code, AI-powered business communication, AI for content creators, generative video app, virtual brand building, single sign-in AI, AI avatar industry, Sora launch strategy, user access limits AI, AI creative tool, AI business advantage.
Podcast Transcript
Jordan Wilson [00:00:14]:
OpenAI has just released Sora two, its newest AI video model. And the conversation so far has largely been centered around three things. One, how good the model is. It's really good. Two, the release strategy, an iOS social media app that was invite only. It seems to put OpenAI in the social media game. And three, the societal impacts of cyclical AI brain rot in the disinformation that it could spread. But that's not what you're here for.
Jordan Wilson [00:00:50]:
And I think those three things aren't where you should be focused either, at least when it comes to Sora two. And I think you need to be focused on Sora two. Here's why. With Sora two, OpenAI has quietly created the world's best AI avatars that I think busy business leaders should be using around the clock. If you can access or at the right way, we're going to be going over that and a whole lot more today on everyday AI. What's going on y'all welcome to Everyday AI. This is your daily livestream podcast and free daily newsletter helping everyday business leaders like you and me not just keep up with what's new in AI because what good is it with just keeping up? You have to actually learn how to leverage it. So that's what we do here.
Jordan Wilson [00:01:37]:
It starts with the unedited, unscripted livestream podcast. But if you wanna take it to the next level, you leverage it by going to our website at youreverydayai.com. Make sure you go sign up for our free daily newsletter. We're gonna be recapping the highlights of today's show as well as keeping you up to date with all of the latest AI news that's happening right now. So speak of speaking of what's happening right now, the reaction to Sora two has been strong. All right. I could have done a episode yesterday on Sora two has been out for about almost forty some hours by now, but I wanted to wait and actually see how people were reacting to this. Cause I can go out and, you know, show you a bunch of videos and, oh, here's some fun things you can create.
Jordan Wilson [00:02:21]:
And I'll maybe just do one or two things to show you that. But I think there's so much business utility here. That's getting lost in all of these other conversations around copyright and the legal headache, around the AI brain rot concerns and doom scrolling, and the societal impacts that technically this Sora app could have. And I think we can't underestimate that. Also make sure stick around to the end. The other rollout part that I kind of referenced, it's an iOS app, but it's invite only. All right. So, we gave away, or we're giving away a code, an invite code today to Sora in today's newsletter, and we're gonna be giving one away, tomorrow.
Jordan Wilson [00:03:08]:
So stick around to the end of the show, and I will tell you how you can enter to get an invite code and to get started in Sora. Because right now, even if you're on a paid plan, a highest tier to pay plan, you still can get access to Sora without that invite code. All right. So on today's show, here's what we're going to go over. I'm going to tell you exactly how Sora two works, the pros and the cons. I'm going to share the details of the texts, the danger, and the legal headache, and it's going to be a big one. And I'm going to tell you how you can unlock, how your business can unlock, this thing and turn it into a viral video app in a legit brand builder for your company. How to look past all the, TikTokification, in the AIification of everything that people are using Sora for right now, And yet actually turn it into a business building tool that I think in the short term could be an unfair advantage for your company.
Jordan Wilson [00:04:09]:
All right. So what the heck is the new Sora? Well, the original Sora was teased like a year and a half ago, and we had to wait like ten months for the time that OpenAI, kind of initially teased Sora until they rolled it out to the general public. And again, it was only for, the highest paid, the highest tiered paid plan originally. But, you know, I think Sora, the original version was really good, but it had a lot of very noticeable fall, faults. So the new Sora or Sora two OpenAI build it as the way to turn your ideas into hyper real videos with sound. So here at a glance is what's new and what's important with Sora two. So the tech is outstanding. It's by far the best AI video model.
Jordan Wilson [00:05:00]:
And I would say nothing else is close, at least right now. Right? We'll wait to see what, you know, Google does with, Gemini three and presumably VO four, their, AI video model right now. V o three was previously, I think, the best, all around AI video model. So we'll see how Google responds. I think they'll respond, sooner rather than later. But the tech right now, it just creates hyper realistic videos from text with physics aware simulation and synced audio music and dialogue. That is the big, upside and the big improvement, at least from Sora two. Sora two didn't have that.
Jordan Wilson [00:05:38]:
Or sorry, Sora one did not have that. Sora two does have, this synced audio music and dialogue, and the original Sora didn't understand physics that well. You know, I think at the time, you know, you could generate, you know, the same text prompt maybe like four, five, six times, and you could get one that had, kind of realistic physics. Right? Someone walking, some type of action on the screen, but it usually took at least a handful of prompts, which took a decent amount of time. And then even after that, you were still just left with, okay, this is okay. Now essentially you get one shot, maybe two, and you get physics that look absolutely real. All right. Right now it's invite only, like I told you, and it is available via a TikTok style iOS app and web portal for invited users only.
Jordan Wilson [00:06:31]:
And you don't even have to be on a paid plan. All you need is that Sora invite code, and you can start generating right now. The tiers are, or sorry, the limits are going kind of up and down, OpenAI said as they let more people in. As of yesterday, it was about 50 generations a day, which is actually a decent amount. I saw someone on Twitter do the math, you know, for, what v o two would charge for the same thing. Right? We don't know pricing ultimately, for Sora two, because it's free. So for v o three, I think someone said, that 50 generations was, you know, a couple $100 a day per user. Right? So this is presumably a very expensive kind of free launch for OpenAI.
Jordan Wilson [00:07:14]:
Right? Everyone's wondering like, well, why are they, you know, reportedly losing money, and why are their GPUs melting? And, you know, why did they live last week literally, enter into three separate multibillion partnerships with different companies? Well, this is why. So some of the key features aside from some of the audio is a cameo feature, and we're gonna talk about that more soon. Essentially, you and I can go in there and create a video avatar version of ourself that is better than probably any of the big AI avatar companies out there, which says something. Right? Because these companies have been working solely on being AI avatar companies and now source in a different category, right? So that's something that you couldn't previously do, with VO three, from Google. And you can do that now with Sora two's cameo feature. And we'll go over some of the protections that OpenAI put in place to make sure that that's not, misused. Right? So you have control when you create a cameo of yourself, and only you can create one of yourself. So you have control over how it's used, and some notifications of when it's used.
Jordan Wilson [00:08:24]:
And then there's also a remix feature inside the, Sora app. So, if you see a generation that you like, you can go in and click the remix button and change a couple of things with the text prompts, and then you can kind of build off or fork other people's AI videos. Copyright issues are abundant. They're everywhere. And I don't think OpenAI is gonna shy away from that in the long term, and we'll talk a little bit more about that as well. But in there what OpenAI is trying to do is they're trying to do an opt out training policy, which, we'll see how that works and more on that here in a minute. But also the deep fake risks are big. Right? So the Cameo thing does require verification, and I think, that's where a lot of people are gonna be focused on, and where I think a lot of the attention in the, the viral nature of Sora two is from that Cameo.
Jordan Wilson [00:09:14]:
Right? You can go in there and you can create a video of Sam Altman, the CEO of OpenAI, doing anything. Right? Think half of the videos that were trending on the Sora app in the first twenty four hours were just of Sam Altman saying crazy things, doing crazy things. Right? And it looks like him. And, it doesn't sound always entirely like him, but some of them do sound like Sam Ullman. Alright. So let's talk a little bit more. So Cameo does let verified users insert their actual likeness invoice into any AI generated scene. Right.
Jordan Wilson [00:09:47]:
Which is pretty cool. And then, like I said, the remix allows users to fork and edit other people's videos, in the pro tier. So even though this is free to use, there are limits and I'm sure the limits are gonna change a little bit, as OpenAI, I'm assuming they're, you know, onboarding tens of thousands of people each day, and it's gonna get up to, millions and eventually tens of millions of users. I would, I would think on Sora, the tiers are gonna change. But we did see that the pro tier. So if you do have a chat GPT pro account, that's $200 a month, you will eventually get twenty second videos. So that's a don't worry about that. The thing on my screen that says watermark free outputs, that's not a thing.
Jordan Wilson [00:10:29]:
That's an error on me. Human midair. All right. So how can you get access to it? Well, here's the tricky part. So not only is it invite only, but you have to sign up with the Sora app. So it's an iOS app. You're not using your, you know, chatgpt.com. You have to get a Sora app and then you use, your chatgpt credentials to log in.
Jordan Wilson [00:10:55]:
Pretty big thing here. Another thing no one's talking about, the chatgpt sign it, which I've been talking about since, the 2025. How I think ultimately, you know, kind of how you have the single sign in for things like Google. Right? You can sign into a thousand services with your Google credentials. You know, so this is kind of the first big time, the first big instance that OpenAI has rolled this out with Sora. So you can log in, with Sora and you actually have to log in to Sora with your Chad GBT credentials. That does, however, keep this in mind that ties them. And I, I did see some other people talking about this in law online.
Jordan Wilson [00:11:32]:
If you sign up for, If you sign up for Sora too, you go into the Sora app, you're connecting those two. And then if you want to leave Sora, maybe it's by design, maybe it's a bug right now, but it will also delete your chat GBT account. So you can't really separate the two. So keep that in mind, right? If you wanna go on Sora and then all of a sudden you're like, oh my gosh, I'm spending, you know, an hour every single day on this app, just doom scrolling through all these AI videos. Let me delete, my account. It can't right now because it'll also delete your link to Chad GBT account. So right now the free tier can generate videos up to ten seconds, but like I said, the chat GBT pro tier, will, reportedly have up to twenty seconds of generation with faster generation and extended capabilities, although that is not out yet. Alright.
Jordan Wilson [00:12:25]:
Let's take a look live. Alright. So, live stream audience, let me know if you can see in here, and also, podcast audience. I'm gonna, be showing a couple of the demos that OpenAI did. So, if you want, you can always go to our website to watch the video version of this, at youreverydayai.com. Alright. But, I don't think you're gonna be missing well, maybe you haven't seen these videos. I'm sure a lot of you have, but they're pretty impressive.
Jordan Wilson [00:12:51]:
Okay. So these are some of the demo videos on OpenAI's website, and they do have sound as well. So I do wanna watch just a couple of them, three of them. And I want to denote not just the, the aesthetics that I think are amazing. Right? Not just from a quality perspective. Also a little background on myself. I used to shoot photo and video professionally, right? For a couple of different roles, probably for eight years. So I've shot thousands of hours of video.
Jordan Wilson [00:13:25]:
I've edited video for tens of thousands of hours. I've shot more than a million photos, with, you know, different DSLR cameras. So I understand visual quality, I would say better than 99% of people. And I've done it at a fairly high level. The visuals on this, the actual visual visuals, the aesthetic quality. So good. Outstanding. You know, I'm not the one to talk about, you know, physics, you know, motion, all of these things, but from a visual perspective, the physics look great.
Jordan Wilson [00:14:01]:
Right? Whereas before, the first version of Sora, not that good. V o three, pretty good. Right? But you can get a mixed bag. Sora, the physics look outstanding to the average person. Alright. So, we're gonna look at the first one. So this first video, it is someone doing three consecutive backflips. Let's take a look and listen.
Jordan Wilson [00:14:22]:
Check. Oh, the landing. So, that is so good. So someone doing three backflips between each backflip, kinda like a little stumble, a little hesitation, which looks natural. And then when this person is done doing the three back flips, kind of stumbling and walking away excited, I mean, it looks so real. Alright. The next one which we have to talk about is gymnastics. And one of the reasons why, this was one of those it was like a a nightmare fuel of AI video generators, Sora one, v o two, and some v o three.
Jordan Wilson [00:14:59]:
Right? Everyone said, hey. Video AGI will be when, someone can get gymnastics right. So here we have a gymnast, on the beam doing a couple of moves. If you look at it closely enough, you'll be like, okay. Well, in theory, she maybe could have fallen there. Right? So the physics aren't maybe, a 100%. I would say they're maybe, I don't know, 80 to 90%. But if you're watching this, unless you're looking for it, it just looks like, oh, this gymnast recovered.
Jordan Wilson [00:15:29]:
Right? She's doing, I I don't know the gymnast moves like a a a flip and a spin at the same time and then a back handspring, right, and then, almost falls, but then recovers her balance. So let's listen. So you can even hear her feet, hitting, it's great. And people reacting, to that in real time. And then the last one I'm gonna watch it is someone jumping off of a diving board into the water, doing a little bit of, of a cannonball there. And it's really good, really good. I mean, the physics on the water, the way the splash, hits the water, and also the way it ripples out. So you have the initial splash and then you have a slow, perfect ripple.
Jordan Wilson [00:16:23]:
So let's go ahead and, we'll watch that and listen to it as well. That's cool. Animal. That's cold. I mean, really impressive stuff. I will look, we will look at one more. This one's alright. So let's, let's watch April 1.
Jordan Wilson [00:16:54]:
So really, really good. It looks like they're doing a a kick flip of some sort, and then they're trying to grind on a rail box. I'm not like a skateboarder, but I think that's kind of what that is. Right. But again, the physics look great. This this person sticks, and it also has that kind of, you know, early 2000 skateboard, like fish eye look. So it looks like a real skateboard video. You know, they stick the first kind of, kick flip.
Jordan Wilson [00:17:24]:
The second one, they're trying to get onto the the rail box and grind, and they kinda fall off, and the fall looks extremely realistic. The sound is perfectly lined up. So let's listen one more time to that one. Even at the end, right, it looks like a young a young kid kind of, you know, yelling at the end. Again, the soundscapes, the motion, the realism, it is really good, which actually is really bad for society. And we'll talk about that a little bit more later. While I'm sharing my screen, I will show, one other thing here. Kind of how you can go in and use people's, kind of cameos, is is kind of what they're called.
Jordan Wilson [00:18:20]:
So again, I'm in the, let me go ahead and share this tab of my screen. Give me a second here. Get get a little bit of an error. Obviously, that happens sometimes. So you can go down. So now I'm I'm on, sora.com. So there's here's all these different cameos that you can use. I don't really know any of these people, except Sam Altman, and there's a couple other, OpenAI members, that I know.
Jordan Wilson [00:18:49]:
So you can literally just click on them and now I can type and create anything of Sam Altman. And that's what I did. I was doing some demos. Right. So here's a couple of things because yes, Sora has this watermark. Right. But on the web version, you can create, not just the vertical, the vertical versions, which you can do, on the iOS app, but you can create landscape or 16 by nine. Right? So, a lot of times, they will create a 16 by nine.
Jordan Wilson [00:19:16]:
Sora will create a 16 by nine video, box, but a vertical actual video. So then in this example, like I showed, it's actually very easy if I wanted to, to cut the, the watermarks out. Because the watermarks, I mean, they kind of go in and out and they move, they pop up in different places. I think OpenAI did that intentionally, as to make it hard, to get around. Right? So it's like you can't just easily crop this watermark except if you do it this way that I did. Right? So I did Sam Altman here essentially doing an ad for everyday AI, and I did something fun that I hadn't really seen anyone anyone else do. I added kind of, you know, viral video style captions, and it actually did an okay job. It didn't get it perfect, but I didn't even know that Sora was capable to do that.
Jordan Wilson [00:20:11]:
It literally puts what he's saying on the screen. So here is, Sam Altman. Let's get the the the the right one here. Doing an ad for everyday AI. So, see if it sounds like him, looks like him. Here we go.
YouTube video [00:20:26]:
We're actually using AI and work in real life. New episodes drop every weekday. So if you wanna stay in the loop, hit subscribe anywhere you listen. See you there. Today AI podcast is kinda my go to way to keep up with how people are actually using AI in work in real life. New episodes drop every weekday. So if you wanna stay in the loop, hit subscribe anywhere you listen.
Jordan Wilson [00:20:42]:
See you there. Alright. So, pretty impressive. Right? But like I said, actually kind of problematic because the Sora, logo is just going around the black box. So I could, in theory, cut this, crop it, and I would have a vertical video with no watermark. There is a different watermarking technology, but it's technically easy to get around that as well. But what I will say, one thing that Sora does really well, soundscaping, we're gonna get into some of the details on this, but, also just with these cuts, you know, it's, it's, they're kind of like these cuts as if someone was editing, from, you know, different camera angles, putting different things together, which you don't really see a lot in a lot of AI video products. Most of the time, you just kinda get, like, one continuous scene.
Jordan Wilson [00:21:30]:
So this does look like a little bit more professional. And I've I've done a couple of generations where I'll get even in a ten second clip, I'll get three to four different camera angles, and it does bring a different level of realism, and it doesn't feel really AI generated. But, I mean, everything from a aesthetics perspective, the composition, the lighting right in this case, you know, Sam Altman walking, along the Chicago, river walk the sun setting it's, you know, hitting perfectly, you know, on one side of his face, right? It's top notch. And it's actually hard for me to believe that we are at this level right now in AI video. All right. Let's keep this thing going. I wanna go over a little bit more of context and what's next and, talk about what I think you should be ultimately focusing on. So opening, I did describe this as kind of the GBD five 3.5 moment, for video.
Jordan Wilson [00:22:29]:
And they're saying it is a generally, generational leap, and I would agree. I do think when Sora one came out, it was the best AI video model, but not for long, right? There was a lot of, AI video models from China and then Google's, VO three just wiped it. But now I would say this is a leap and it is extremely impressive. And, not only that, but just from a world simulation perspective with the physics, extremely impressive. Also its capabilities immediately raise fundamental questions about creativity, copyright, and authenticity. That is the downside about this huge unforeseen, honestly, generational leap, in AI video quality. I don't think the world's ready for it. And I think too many people are going to be focusing on the social media aspect, the TikTokification of AI.
Jordan Wilson [00:23:27]:
I think that's what most people are going to be looking at. The thing I'm honestly worried about is the deception side, the misinformation and disinformation side is going to run rampant. All right. Especially right. Even, until we had VO three, we didn't really have any AI video models that could match up, not just audio music, background noises, complete soundscapes, right. That sound freakishly realistic. So let's talk a little bit more about the physics aware video that really gets it right. So, OpenAI has what they're saying is improved temporal consistency.
Jordan Wilson [00:24:13]:
So So that's essentially from shot to shot scene to scene, these multiple scenes that are stitched together. It's maintains the kind of complex lighting shadows and character movements across scenes. So I did a couple where I had, you know, oh, Sam Altman walking through a busy crowd and it would shoot, jump to a different, angle. And, you know, even if there was ten, fifteen, 20 people there, it kept the same consistency, from these two different, cuts within the same, video output, which is crazy. Right? Not only the same character consistency, but the same position consistency. Right? It's like, oh, there's, you know, the person in the red shirt is behind, the guy in the black beanie in shot one, and then the same thing in shot two, but it's clearly a second later, and they're moving in the same direction at the same pace in the same position. Or one person was walking faster where clearly they would have eventually overlap them. But the first shot cuts to the second, and then by the third shot, that person overlaps them.
Jordan Wilson [00:25:16]:
It is. Again, I'm at a loss for words how impressive the physics are. Same thing. You know, there was, a demo video that open AI open AI did where a basketball kind of hits off the backboard very realistically. I, I played the, the water, splash and how it didn't just get the splash in the angle that it would, the, the water when someone jumps into a pool, that's not moving the angle of the splash, but also the correct ripple in timing, that comes after. And the model, I think demonstrates an understanding of cause and effect that objects interact with environments in physically plausible ways where I don't think we had this six months ago. We definitely didn't have this with the other version of Sora. Like I said, a lot of times you'd have to do five, six, seven, eight, iterations of the same text to video prompt in the original Sora just to get something that might be passable or you're like, okay, well, you can't notice the bad physics that much in this, or you could edit around this.
Jordan Wilson [00:26:15]:
Now it's like I said, essentially one shot in physics look perfect, unless you are a lifelong video editor or a physics professional. The other one, the native audio integration. So a lot of people use to share these workflows online, you know, taking the original Sora or some of the other AI video, tools back when there was no voice sync. And, you know, people would use six, seven different AI tools and you got something that looked really bad. That's why I never really did anything like that on this, on this podcast because I'm like, ultimately, it looks bad. But now it can generate synchronized dialogue, ambient sound, music, and sound effects. Literally, in one pass, it is a complete soundscape. And the audio fidelity captures environmental nuances like wind noise over voices and lecture hall echo effects.
Jordan Wilson [00:27:13]:
It's nuttier like I sometimes say, nuttier than a squirrel on keto. And also it can again, the bad stuff. Well, right now it's also replicating audio that's copyrighted, like, you know, Peter Griffin, you know, from Family Guy, Eric Cartman, from South Park, SpongeBob with spot on accuracy. Yeah. There's gonna be a lot of legal fights happening over Sora. People are just gonna be, like, oh, why did this show get canceled? Let me just go ahead and recreate a whole new season on my own. So let's talk about that. The copyright problem and open AI essentially announced, Hey, you know, publishing companies, media companies, big, conglomerates, small creators, whoever it's on you to prove right.
Jordan Wilson [00:28:03]:
Even though it's not technically how copyright law works. OpenAI is literally an actively trying to rewrite the rules of copyright law here in The U S granted copyright law here in The U S is a little antiquated. I'm not saying I agree with it or disagree with it. It's antiquated for sure. It's most of it. Well, all of it, I think was pre internet, pre digital media. Right? So copyright law is very old. It's pre internet.
Jordan Wilson [00:28:31]:
And they're using it by default, and you can go on there in the Sora feed and see SpongeBob. And, let's see what else. I'm gonna I'm gonna just, scroll back here, to the main page if it if it loads for me. I'm gonna show you what I see here. So, it they have a for you section. Right? I'm not even gonna show it, on my, on my screen here. I don't I don't wanna accidentally be breaking, but, I mean, you have, Pikachu. You have, Family Guy, Peter Griffin.
Jordan Wilson [00:29:07]:
You have SpongeBob. Lot of SpongeBob. People are going nuts with a SpongeBob creating their own. You have Ronald McDonald, more pico Pikachu, Rick Rick and Morty. Is that what it's called? Some ton of Nintendo, references. Mario. Right? So many of these. Everything is just there.
Jordan Wilson [00:29:25]:
So, OpenAI is though lobbying governments, claiming that AI training is fair use for American competitiveness and national security. So OpenAI isn't necessarily concern seemingly with copyrighted material. And right now, what they're essentially saying is, hey, you will need to report each specific violation individually. Right? So, you know, these companies can't reach out and say, okay, well, you can't use Pikachu. No open AI says, go find Pikachu and then report it to us each and every one. Again, not normally how copyright works. So open AI is just kind of trying to rewrite the rule of copyright. This is obviously gonna be challenged to the courts, I would say, pretty quickly.
Jordan Wilson [00:30:12]:
And you might be wondering, like, okay, Jordan, why are you, you know, you know, picking on, OpenAI and Sora here with AI video copyright? Why aren't you doing the same with Google in v o three? Because people are obviously doing this with v o three. And again, sometimes you're gonna get a response like, oh, we can't create that. But clearly, ultimately, people are creating it. Well, v o three from Google, they don't have a social media network. Right? So there's no scene of the crime, so to speak. OpenAI, I think the Sora iOS app is going to be the number one app, The number one free app on the iOS app charts. I think it got up to number five. I think it's gonna get up to number one.
Jordan Wilson [00:30:56]:
I think it's gonna be wildly popular. The viral factor is through the roof, but that is the scene of the crime. Okay? I I don't see how there's not massive, massive lawsuits, that are gonna be hitting OpenAI with Sora. So let's talk a little bit more about the cameo and the deep fake. We're gonna move on from copyright infringement to deep fakes. So the cameo feature, I can go on and create a cameo of myself. I was still on the fence about that. I think I'm going to, I don't know if I'm going to write, but you can, so how it works, it's a one time verification.
Jordan Wilson [00:31:35]:
You have to do it with your phone because like I said, you have to sign up at least for now. You have to sign up, for Sora on the iOS app. So you have to do it on your phone. So you essentially have to sign up to this, TikTokification of an AI slot machine. But then you can, after that use it on the web. Right. So I'm not going to be using Sora on my phone and doom scrolling. I'm not a social media person.
Jordan Wilson [00:31:59]:
So inside the app and you can only do the cameo in the app, but it's a one time verification and you essentially have to say three random numbers and then you turn your head, kind of back and forth, and then you have a cameo and it listens to your voice and duplicates your voice, everything. And they're pretty impressive. I've seen, you know, Cameos, like I said, the same all in one looks really good. So users though, once you do create a Cameo, you control permission levels, and then you also receive notifications when your likeness appears, and you can revoke access or remove videos anytime. So you can create a Cameo of yourself, and say, hey. Only I can use it. No one else can use it. So you do have control kind of over your AI avatar or your deep fake.
Jordan Wilson [00:32:40]:
And then there's other, sharing options. You can only do it with certain people. You can only do it with mutuals, or or you can do it with everyone. So as an example, you know, Sam Altman just put his out there for the world to do. So anyone can go make Sam Altman say anything, do anything. There is one that kinda went viral on all social media. Right? It showed, you know, Sam Altman, you know, stealing GPU's from target. Right.
Jordan Wilson [00:33:05]:
So, but that's the cameo feature and more on that here in a little bit, because I think that one's important. The watermarking system. So I already showed you one way, you know, I think that the watermarking is clever. It's an animated watermark with the Sora logo. It pops up and moves around the video. So I don't think most people are going to try to edit it out, but I do know that there's going to be, apps that literally you're just gonna upload your Sora video and it's gonna clean it out. That's gonna happen. Right? So the the visible watermark is easily defeated by screen recording, you know, all those things, but it also not just the watermark.
Jordan Wilson [00:33:43]:
It also has the invisible, CPT, C P two A, metadata. So that's the Coalition for Content, Provenance and Authenticity. That as far as I know is also easily easy to get around. Essentially there's apps online or you can just do a screen record, and that is good chance you can get rid of that too. So although OpenAI and the other AI video companies are, I think are doing their best. It's not hard to get around this. It's not hard to get around that. So you have to think of, okay, we have to be prepared to live in a world of, well, nothing's real and everything's fake.
Jordan Wilson [00:34:24]:
Everything's AI. Right? I'll probably do a dedicated show on that later, but you, you, you kind of have to know that simple editing techniques will get rid of that watermark. And I mean, talk about things like court evidence, video evidence is it's not going to be a thing. Right? Cause yeah, this is Sora two. Imagine Sora three, imagine Sora four. It's going to look more realistic than real life. Right? So yeah, it's the societal changes, are going to have to be huge to keep up with this. Right? I think we had as a society decades to adapt and understand like, Oh, something can be photoshopped.
Jordan Wilson [00:35:05]:
Right? A photo doesn't necessarily mean that everything in that photo happened. But I think we had decades to acclimate to that and understand what that meant and the repercussions of that. Video were, were there. We don't get decades. We don't even get a full year. We got a couple of weeks. Video. Just assume all video is fake.
Jordan Wilson [00:35:30]:
Doomscrolling. Let's get to more concerns. We talked about the cool stuff. Think we're gonna skip over the bad stuff. I think let me be honest. I think social media is bad. Right? You might be saying, oh, Jordan, you're on you're on Twitter. You know, you're on LinkedIn.
Jordan Wilson [00:35:47]:
You do the LinkedIn live stream. Yeah. I don't sit there and scroll it for hours. Certainly don't doom scroll. You know, I do read, content on Twitter. I actually like the, yeah. All you people that say, oh, I hate Grok. No.
Jordan Wilson [00:36:00]:
I hate using I would never recommend any business use Grok as a large language model, but I love going on Twitter and reading a long thread on, you you know, something on the latest, you know, AI research, and then interacting with Grok and talking about it and understanding it better. So anytime I'm on social media, I'm usually actually learning, right, and then I screenshot things for myself to go back and, read later. But the doom scrolling thing is gonna be huge. Right? So you open this thing, you open Sora on your phone, and it's, you know, Sam Altman's head spinning in a toilet and, you know, all these crazy things that you never thought possible. So I won't say that this is like the meta vibes level of, trough or AI slop. But it is definitely a sloppification, of the real world, right, where I think the MetaVibes AI app that they're releasing, it's it's gonna be much more, you know, artistic, things that look overly AI. That's not Sora two. That's not the Sora app.
Jordan Wilson [00:37:03]:
Most I mean, yes. Obviously, some things are intentionally supposed to make supposed to look fake. This looks more like TikTok. Right? This looks like something you open up, and it's real people talking in real environments, interacting with each other, people you might actually know and, quote, unquote, follow online anyways. But there are, OpenAI, to their credit, has actually built in and are trying to build in some safeguards. So, yes, it's very swipeable. It is doom scroll central. So, there's the for you recommendation algorithm, that mimics TikTok's addictive engagement patterns.
Jordan Wilson [00:37:38]:
But adult users do get nudges to take breaks if they've been on it too long. Teen accounts can have mandatory cool downs, and there's parental controls and opt out personalized feeds. But ultimately the the platform is biased toward rewarding creation and remixing over just passive consumption. So that's what OpenAI is saying they're doing. They want people to create. They don't just want people to consume. I do believe parts of that. Right? Any other company that says that it's like, ah, I don't believe it.
Jordan Wilson [00:38:09]:
Any other company that's making the social media, OpenAI, I kind of believe it. You know, so we'll see if they stick to that. So just goes to beg the question, who the heck is this thing for? Well, I think social media creators and influencers that need a low friction short form content generation for TikTok, YouTube shorts, they're going to use this, but also I'm being honest. I think TikTok and YouTube shorts might lose a little bit in popularity. Right. Which I don't, I don't know if it's a good thing or bad thing. Think of like TikTok, right? I'm not on it, but you know, sometimes literally I have friends my age and sometimes a little younger that will record TikToks, like screen record them on their iPhone and send it to me. Right? Because if someone sends me a TikTok, I'm like, you know, I'm not watching that.
Jordan Wilson [00:39:01]:
Right? So, so I know what, you know, I know the overall, you know, what a TikTok looks like, how it works. Right? But people who are on TikTok and these other, you know, short form doom scrolling platforms, you know, reels, there. I mean, real talented people and it takes a lot of work and a lot of effort sometimes to create especially high engagement, high quality TikTok videos, reels, etcetera. That's gone. You can create very impressive, you know, TikTok style videos of you doing anything with any talent with zero effort. So I do ultimately think the TikTok, the TikToks and the reels of the world are gonna be in trouble from this. And just, yes, those, those platforms do have policies on AI generated video. Again, people are always going to be two steps ahead.
Jordan Wilson [00:39:59]:
Every, everything that is produced on Sora is ultimately going to be passed off to every other social media network. So also another thing to keep in mind, this is coming out to the API as well. Right? So there's going to be, eventually probably hundreds and then thousands of other platforms that are using this technology, not just the Sora app. So one thing OpenAI said is that they want it's they're kind of doing this six month test. You know, they stated that their goal is helping users achieve true goals like fitness, social connecting, social connection, or starting businesses. And their core success metric is whether users feel their life improved over six months compared to not using Sora. And if that metric fails, OpenAI commits to significant changes, quote unquote, or discontinuing service entirely versus just accepting the doom scrolling reality. Alright.
Jordan Wilson [00:40:56]:
Well, I'd hate to tell you this OpenAI. You might get a couple nice stories. People, yeah, I feel better. I accomplished some goals, but ninety nine percent of people are just gonna be doom scrolling this thing, wasting time. This is gonna make anyone's life better. Not not the whole yeah. You're gonna get a couple nice use cases. So we might see a nice feel good video, you know, from OpenAI saying, alright.
Jordan Wilson [00:41:19]:
Here, Jimmy Jimmy, used Soarer and got motivated about fitness. Sure. Cool. No, this is a doom scrolling machine. This is AI generated AI slop. Right? But there's your business can use slop. Let's end it on this. What is the big takeaway? I just want on a random, hot take Tuesday, not on a Tuesday.
Jordan Wilson [00:41:42]:
Well, is it TikTok Brain Wrap? Yes. Can your company take advantage? Yes. Here's how. Sora.com is going to be the cheat code. Okay. So yeah, you gotta do the iOS app. You gotta get the app first. You gotta get the app to make the cameo.
Jordan Wilson [00:42:02]:
Right. But from there, huge. Okay. What I see and how I could see this really rolling out. And if you are a, an AI native company out there looking to scale, with limited resources, or if you're a big enterprise company that wants to push innovation, this is what you should do. Your CEO and leadership team should be using these cameo, for everything. Right? And you might be saying, okay, even if you're on the pro plan, what's a twenty second little video gonna do of our CEO. Okay.
Jordan Wilson [00:42:36]:
Well, have your CEO sit down, record a couple hours of video. Right? So actually him or her speaking. Then from there, you have this Sora machine. And if they go in and create a cameo, you have countless countless marketing angles, countless, presentation styles. Right? Like I'm even thinking for myself. Yeah. I can get on video at any time. The editing takes forever.
Jordan Wilson [00:43:09]:
Right? If I wanna do different scenes, if I wanna, you know, be outside at, you know, Chicago or, you know, walking through a a group of people, whatever, that takes time, effort, money, budget, things that, you know, a lot of people and a lot of companies don't necessarily want to invest. You can sit anyone down and do a talking head video, very easy, but Sora can do everything else. Right. And I think there, there is a little bit of inspiration to be had, right. Scrolling through and seeing what other people are doing, but then think about it from your business perspective. But I think ultimately the cameo can be your company's secret weapon. Maybe your CEO is too busy. Maybe your VP of training can't do everything at all times.
Jordan Wilson [00:43:56]:
Right? So having the cameo and being able to control who has access to it, who can use it is I think such a huge cheat code for business. This is why the AI avatar industry has become an industry that is generating billions of dollars in revenue because so many people are using it to train. So yes, that's a little different because you can do, you know, hours of talking head videos, right? You can still do talking head style videos in here, but in short bursts, but you can put your CEO anywhere on any stage talking to anyone, demoing any product, Right? Traveling, visiting, you know, visiting, quote, unquote, visiting potential clients, in in a new area of the world that you're trying to expand to. This is literally like having an entire film editing crew and a teleportation machine all in one, and it's instant. So you do have to think a little create, a little creatively and outside of the box. But I do think the big cheat code for your business. Don't use the app. Don't use the Sora iOS app aside from, to sign up and to create the cameo.
Jordan Wilson [00:45:09]:
But then from there, possibilities are endless. All right. I hope this is helpful. I know this went on a little bit longer, than some of our normal episodes. So thanks for sticking around. And I did tell you, you're like, okay, this sounds great. I want to do it, but I can't get access to Sora. Okay.
Jordan Wilson [00:45:24]:
Well, go find this episode. So if you're listening on the podcast, in the show notes, I always put the link back to the LinkedIn live stream for this very episode. Just go repost this. Alright. And then we're gonna be giving away a code of Sora in our newsletter tomorrow. So you got, like, twenty hours. Go repost this show on LinkedIn, and we're gonna be giving away a Sora code in our email newsletter tomorrow. So you gotta open it.
Jordan Wilson [00:45:51]:
You gotta read it to see if you won. So thank you for tuning in. If you haven't already, please go to your everydayai.com. Sign up for the free daily newsletter. Thanks for tuning in. We'll see you back tomorrow and everyday for more everyday AI. Thanks, y'all.