In short
The episode argues “Hollywood is upset about Quilty” but claims the underlying idea isn’t new: AI “script coverage” and box-office/viability scoring resemble earlier tools and even the NFT scam pattern—wrapping existing data/LLM workflows in a new product.
Guests
Addy (co-host; discusses testing AI image/video tools and practical workflows) and Joey (co-host; contributes technical commentary on identity models, uncanny-valley detection, and video/image quality issues). Key topics: Foda Labs’ identity-focused human/pet likeness model (30–50 reference images via a LoRA-like identity layer; strong lighting/skin texture but fails on head shape; uncanny-valley limits around ~300 familiar faces). Fotelo for real-estate “re-photographing” (HDR/exposure/rectilinear corrections; not full virtual staging). Wand 2.7 for synthetic character creation, text rendering, color control, multi-image editing, and sequential storytelling. Pixverse V6 and LTX 2.3 updates (quality/resolution gains but nonsensical actions; LTX adds “reasoning”/emotion restraint). Quilty Score: claims to evaluate story/craft, commercial viability, cultural resonance, and production reality; hosts counter that predicting hits is inherently unknowable.
Notable examples
Artemis launch filming with Apple immersive cameras; “Zillow gone wild” AI staging; House of Cards/Housemaid as data-driven but still unpredictable outcomes; CloudCode/GitHub leaks and malware; an Anthropic internal-email blackmail anecdote.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOApril Fool's Day Reflections
0:45 to 2:15
Discussion about the impact of April Fool's and online trust issues.
“Like I thought when I saw that article, I thought Sam Altman is crazy enough to build a hardware camera.”
New AI Models: Foda Labs Introduction
2:15 to 4:17
Introduction to Foda Labs and its focus on human likenesses.
“I gave it some screen grabs from us in the past.”
Exploring Foda Labs Features
4:17 to 7:20
Detailed discussion on Foda Labs' capabilities and performance with images.
“This was a straight text-to-image to make, to just use it as, like, a photo-generating model.”
Real-World Applications of AI Models
7:20 to 8:49
Exploration of potential applications for Foda Labs in various industries.
“that it's like, okay, we're going to take the foundation model and make this really good at a specific thing of human likeness.”
Real Estate Photography and AI
8:49 to 11:28
Discussion on AI's role in real estate photography and staging.
“Like, there is no fall-off in the real distance versus like close distance because generally lenses have like a...”
New AI Wand 2.7 Features
11:28 to 13:00
Overview of the new Wand 2.7 and its advanced features for photo editing.
“Are you familiar with this topic in general?”
Exploring AI in Color Grading and Image Generation
14:02 to 17:04
Learn about advancements in AI tools for color grading and image generation in media.
“Like you design your color board and then align it.”
Evaluating AI Models: Performance and Potential
17:04 to 19:40
Discuss the current state of AI models in generating realistic images and videos.
“right like um the first three months we have seen some new models come out but everything's been just like an incremental increase in quality control, what have you.”
Innovations in AI-Driven Emotional Expressions
19:40 to 22:16
Understand how AI can enhance emotional expression in generated performances.
“I don't even know what he uses as the input.”
The Future of AI in Film Production
22:16 to 25:18
Explore the implications of AI technology on film production and storytelling.
Show all 20 chapters
Cybersecurity Concerns in Tech Companies
25:18 to 27:54
Learn about recent cybersecurity issues affecting tech companies and AI models.
“signage thing that is reading your emotion and then real time interacting back to you Everything you've seen in an AI dystopian movie, it's for that.”
Analyzing AI in Hollywood
28:03 to 28:30
Explore how an AI app analyzes scripts for box office predictions.
“an impressive feed of like, how did you get this great PR for Variety to publish this crazy story?”
Skepticism About AI's Novelty
28:30 to 29:38
Discussing the existing tools in Hollywood and their similarities to new AI models.
“There are a bunch of platforms that already exist that have done this for a while.”
The Limitations of Prediction Models
29:38 to 30:58
Understanding the challenges of predicting box office success using AI.
“and thing that we should probably get into in the future.”
Case Studies of Movie Success
30:58 to 33:16
Analyzing historical data and surprises in film box office performance.
“And then everything was depending on the first box office numbers for that weekend.”
Data-Driven Decisions in Filmmaking
33:16 to 34:25
Exploring how data influences decisions in film production and casting.
“I would imagine so much of their development cycle is data-driven, right?”
The Role of Human Judgment in AI
34:25 to 35:00
Discussing the necessity of human decision-making despite AI advancements.
“That's a terrible, terrible paraphrase of it.”
The Discontinuation of Mac Pro
35:00 to 36:39
Reflecting on the end of the Mac Pro tower and its implications for users.
“Like, it's just that's been the case for a million years already.”
Market Trends in Computing Hardware
36:39 to 37:44
Analyzing how consumer needs and economic conditions affect hardware production.
“And so the product line is kind of changing with it.”
Wrapping Up and Future Considerations
37:44 to 39:12
Concluding thoughts on technology changes and audience engagement.
“Also, the use cases of what you would need that beefy of a Mac Pro tower for are slim.”
Transcript
Automatic transcript. May contain errors.0:00Joey:This is the Tilly Norwood of AI development stories. This is not new. There are a bunch of platforms. This is like the NFT scam all over again.
0:10Joey:All right. Welcome back to Denoised. Addy, how are you doing? I'm doing just fine, sir. How are you? I'm good. April Fool's Day. Can't really trust anything you see online today.
0:20Addy:I got fooled and then you took me out of it. Thank you for that.
0:24Joey:I did have to like double check myself. What was it? It was that Sora. The engine cinema thing from OpenAI. They were going to do their own camera, an open AI camera. Somebody else texted me that too.
0:34Addy:And I was like, beat you to it. I already know it's fake.
0:38Joey:April 1st doesn't hit as hard this year now that I just don't trust anything online anymore. No.
0:45Addy:And the world is upside down anyway. So like it could actually happen.
0:50Joey:It's been like, not today. I don't have time for this. Not today, please.
0:53Addy:Like I thought when I saw that article, I thought Sam Altman is crazy enough to build a hardware camera. Like he really is.
1:00Joey:It was kind of good enough because it was like ridiculous enough and plausible enough where it's like, maybe. But then you're like, no, wait, no, that doesn't make any sense. Okay, so we got a couple new models that kind of popped up to talk about. And yeah, a couple other AI tools that are pissing people off. So let's talk about them. So the first model I want to talk about that kind of popped up on my radar is Foda Labs. And this model is sort of saying, and I'll talk about it's sort of a new model, but I'll talk about that in a second. But this model is targeted specifically at human likenesses, people kind of trying to capture a photographic look, aesthetic, and have really good and accurate people, especially if you give it a reference image of someone being able to keep them intact and alike.
1:49Joey:Character consistency on steroids, maybe? Yeah, I messed with it on a project we're working on, and I gave it some pretty crappy archival photos, and it did a really good job. I can't share those because the project is still under wraps. We're bros, man. Let me see it. But I can share some other tests that I did quickly. So quality here might not be as good as the other test I did, but I gave it also the input images are not that great. I gave it some screen grabs from us in the past. Oh, dear. so these were the input images sure and i had it make some tests that's not you just using that this was not that's kind of me oh this one that's that's pretty no i have more of an oval face that's
2:35Addy:more of a circular face oh yeah yeah that one's close yeah that's actually the miami face joey
2:41Joey:huh that's actually pretty yeah that's not bad i mean i'm impressed by the lighting and skin texture and that kind of stuff the likeness i mean yeah given the input image i had to do my
2:52Addy:advice for both of us yeah that's nice okay that's creepy so this one fell apart so one thing it does
2:58Joey:fall apart it's really designed for one person in an image right now right you can it has sort of its own laura feature so you can give it 30 to 50 images train it on a character and then call that up in the output so it's sort of a laura under the hood and then you could use it to apply to multiple people. This reminds me of
3:18Addy:the headshot generators online. There's a bunch of them.
3:22Joey:Yeah, kind of. Yeah, it would work well for that. These are not the best input images and not the best prompts. This was a quick test, but I found in the other test I did, I was really impressed. Crappy archival image. Good job of making something that looked like that person. Also, given the fact that we don't, we only have one archival image of this person, so we're, it's not like us where it's like, oh, we're real people and we know what we look like.
3:46Addy:Yeah, and I I don't think, Joey, correct me if I'm wrong, but I don't think like a higher resolution reference photo of us would do any better because I think the model itself, you know, the latent space and everything is low resolution. So the control net is downsampling whatever you give it anyway.
4:03Joey:I think it would do better if I these are pretty low res and cut out. And also we have headphones on. Yeah, if it was a cleaner headshot, it would do better.
4:10Addy:Yeah, I think it is messing up our head like head shape. Yeah, that's why it's elongating us on a horizontal direction. Yeah, for sure.
4:17Joey:These were just straight. Well, this was an up-res version. I don't know why I did that. These were just straight text-to-image.
4:24Addy:Of, like, some other person, right? Not you.
4:27Joey:Yeah, not us. This was a straight text-to-image to make, to just use it as, like, a photo-generating model. So it's supposed to kind of capture that, like, DSLR.
4:37Addy:It's, like, 90 % of a DSLR. Look, again, we are so susceptible in the Uncanny Valley region.
4:44Joey:like we can instantly detect a fake person yeah or if it's like off or not quite like yeah
4:50Addy:yeah like um you know photographically everybody knows i think somewhere around um 300 people more or less like that's that's sort of the number where that's our closest social circle and then we can kind of just recollect them by face oh yeah biologically number yeah so like if If any of those 300 people are reproduced with this system, then we'll instantly detect that it's somewhat off. But if it's outside of those 300 people, we won't be able to really tell.
5:20Joey:This one's also interesting because I realized it gave, it's the Artemis, like put us in the Artemis uniforms, which like that's launching today. Yeah.
5:28Addy:Is it manned or unmanned? It is manned. It's crew form. All right, nice.
5:31Joey:It's like launching, like if all goes well, like an hour from now, I think. Should we do a coverage of that on the pod? I mean, are they doing, are these an AI on the spaceship? There's probably zero AI. Actually, I didn't put it in this link, but I did see a link that they were filming the launch with the Apple immersive cameras. I'm assuming it's the Blackmagic ones, but they're filming it for Apple immersive. That makes total sense. Which is cool, because I remember back to the IMAX films of Saturn V, when they shot those launches with the super high speed 35 mil. Maybe it was even higher, maybe it was like 65 mil.
6:05Joey:And like now we need that equivalent today of like the current space launches. Yeah. You need Christopher Nolan directing the launch.
6:12Addy:Can we redo that again? Take two. Like, no, no. It launched. It's gone.
6:19Joey:Oh, yeah. I mean, once this comes out, it will have launched. Okay, so one thing I do want to talk about that came up online was what's behind their model? Because they sort of said, oh, it was like an original foundational model. But then some people are like, what did you actually do? And so they posted a statement. It's a system of multiple models, each optimized for a different part of the generation pipeline. Our proprietary model focuses on identity, which enables us to generate and edit photos of real people and pets. For base image generation, Photo uses leading foundation models, open and closed source, including Nano Banana.
6:56Joey:And then on top of those, we've trained our own identity model on both in-house data and user uploaded photos. That identity layer is the core of Photo's differentiation. So it's an interesting take here. where it's like they didn't build their own foundational model, which honestly makes sense because it's like at this point, that's hard and resource intensive.
7:14Addy:And you're not going to beat the guys that are doing it well.
7:18Joey:Yeah, but they built their own data set on top of it that it's like, okay, we're going to take the foundation model and make this really good at a specific thing of human likeness.
7:26Addy:Yeah, the intellectual property is, I think, the solution layer. So cobbling together the comfy Y workflow, if you will, with an LLM, with Nano Banana, with a VLM, with a style transfer or a DSLR transfer and chaining all that together and then serving it out as a consumable product to the user. That is intellectual property, I think, in my opinion.
7:50Joey:Yeah, I mean, they built this middle layer that does something, the specific thing well. So, I mean, we're going to try this for some of the identity things if we need to deal with stuff with people. It kind of also makes me think of the Ben Affleck thing where it's like if you build that kind of middle layer that is tailored to your film, your look, your whatever, that's still useful, especially in steering and guiding the models for things you need.
8:15Addy:Right, because you're not looking to generate stuff from scratch. You're looking to augment what you already have. And if you have a bunch of footage, a bunch of characters and pictures already, then you can actually extend the life of that beyond what you could normally do.
8:30Joey:These were more text-to-image output tests. They look good. I mean, but honestly, you know, because it's not based on any real person, it's like you could get this out. Yeah, that one is super realistic. Like, if you go back to the last image, the one with Busy Market. The chef or the market?
8:45Addy:Yeah, like the thing that is breaking this is the bokeh on the camera is completely wrong. Like, there is no fall-off in the real distance versus like close distance because generally lenses have like a... Like that this would be more... than like the guy in black. It has like a gradient fall off typically, especially with the larger sensors, you could really see that. And again, the spice, yeah. I mean, it's just too consistent, too uniform. But like these cues instantly tell you like something's off, but you don't really know what. Yeah, that's a good point. All right, yeah, so new model. I think it's worth knowing about.
Read the full transcript
9:25Joey:So who do you think they're targeting? Like who's the clientele here? I mean, I think it's any, like, you know, if we build it in the pipeline of like any of these apps or stuff that are dealing with people that are trying to, you know, make consumer level products that are very face focused, this, you know, could be the good back end for that. I think, oh, yeah, also like one of their kind of other selling points was like if you take a crappy photo of a memory and you want to restore it. So like blurry photos, out of focus photos, far away.
9:55Addy:You want to re-photograph that moment, then put it through.
9:58Joey:Yeah, while preserving not making it look AI. I think that's also been their selling point.
10:03Addy:Okay, along these lines.
10:04Joey:Very consumer-level stuff. I'm not looking at it for consumer applications, but consumer-level stuff.
10:09Addy:If you can pull up a browser and go to Fotelo, F-O-T-E-L-L-O. F-O-T-E-L-L-O. Fotelo, I think that's it. Yeah, that's it.
10:18Joey:So this is along the same line, but instead of people, it's real estate. It's buildings and interiors. Just for staging photos. Yeah.
10:26Addy:So like you can take a like a crappy photo of interior bedroom. Yeah. These are just the people that use it. But let me see if you can find some examples and what Fotelo. Yeah, that's perfect right there. So one of the challenges with real estate photography is exposure, multiple exposure, HDR, and then aligning all the HDRs into a single frame. And then, of course, having rectilinear lensing. so like all the lines appear straight and then correcting all of that takes a lot of time in photoshop or lightroom it's totally doable by somebody like you or me but this takes all of that effort out of it so all you do is take a photo with you know a regular dslr just one and then it'll do the work of like a seven exposure hdri this is cool okay so this is just really
11:14Joey:fixing exposure this is not the virtual staging correct yeah it's just enhancing it's read a
11:21Addy:re-photographing the moment, kind of like what the previous solution did.
11:26Joey:Okay, this is cool. This is very practical. Have you seen those crazy AI staging photos that go off the wall? Are you familiar with this topic in general? Staging homes for selling them? Staging homes for selling them, but using AI to, you take your little real estate photo of the empty room, and then AI adds the furniture and stages it.
11:45Addy:That seems like it's a recipe for
11:47Joey:disaster so that is a very popular use case now with like realtors and doing this ai staging which people are like it really doesn't make it does not it helps it makes it worse to get a sense
11:56Addy:like jeffrey dommer's basement turned into a fancy living room yeah but then this one was like a
12:02Joey:zillow gone wild kind of thing that was um they use the ai staging app on this house that is like an extreme extremely beat up and run down and then it has this like millennial modern furniture staged in the place.
12:16Addy:I mean, this is exactly what a millennial would do because they can't afford nicer homes. This is all they can afford, right?
12:22Joey:So they would just
12:25Addy:upcycle it, as they say. I mean, look, jokes aside, it is doing a good job relighting and repositioning the furniture items. Technically, it is doing an amazing job, but of course, the taste and the creativity there is off.
12:40Joey:Yeah. All right. So I thought that's where your thing makes a lot more sense. That's it. I think you
12:44Addy:combine this with that you can get to somewhere really awesome and then if you've ever dealt with like real estate photographers they're like wedding photographers they charge an arm and a leg like they'll come in for like you know six seven thousand dollars to do a house yeah because
12:59Joey:they would also need to have that specialty to process the photos it does take them a lot of work on their end okay next other ai model wand there's a new wand wand 2.7 i love i love me some one just for photos okay so this one i don't think it could run local it's just cloud service but some of the highlights portrait customization so is that going to be maybe character consistency like thing or design in a character it says for bone structure and eye details to subtle face features create unique and authentic faces so like more like a character design like a fully
13:33Addy:synthetic human that you can really sculpt if you will yeah like a meta ai meta human kind of thing Exactly.
13:40Joey:Advanced text rendering, which is cool. But like, I mean, it's kind of now now we're like, yeah, that's perfect, of course. You better have that. You better have sharp text. Precise color control. That's pretty cool. So you got to give it a palette and have it applied to the image, which would also be great because that's a big issue that's sort of ignored with AI video generation and image generation as you're making the images and consistent characters. but then also how do you make sure the lighting and the vibe of the scene feels the same shot to shot without dealing with that later on and resolve and color if you can get the outputs
14:11Addy:to have like a consistent color palette yeah and you can also maybe do some level of color correction perhaps like work with the color grader to define those color palettes and then go back and recolor your footage to match all of it up i don't know i'm thinking that could be possible yeah yeah
14:28Joey:or other coloring tools in the future that can kind of do that. Like you design your color board and then align it. I think, you know, there's like a color lab or there's like AI coloring tools that plug into DaVinci. That is doing color matching.
14:43Addy:I think Premiere has something like that now built in, AI-powered color matching.
14:47Joey:What else we got? Multi-image editing. So, I mean, that's similar to like C-Dance. Banana, multiple images in. Sequential storytelling. generate up to 12 sequential images with perfect stylistic and that's like consistency right it sounds like it where you get the kind of same images in the same generation which it's good you know we want options and you want to have perfect for storyboarding and stuff like that and then interactive editing use precise box selection to align your creative intent with ai so something that i think the vo or um yeah vo but also nano banana kind of accidentally had where you can annotate an image and then it can follow it i'm wondering how that input control works if you're like using one through like an API or comfy or something else if you have to like make a mask and give it like a matte like your input image and your mask to be like I want this thing here changed I'm curious how that would work on other platforms I mean I always welcome in painting
15:41Addy:because you know people like us we are never happy with the first generation but we do like some aspects of it and then we want to tweak it go in and just take this thing out take this thing in so in painting out painting for a model the better it does the better for us yeah especially
15:55Joey:being able to have that control if you're like oh that's good don't mess the whole thing up this one just just take this one erase all right so i had it also generate some shots i don't know what these shots are like lord of the rings so we're in middle earth this is one two seven this is one two seven with the same i mean this one is so freaking off this one's great definitely not as good as Fota slash Not a Banana, and when you're giving it input images of people. I mean, also, again, these were the same crappy input images.
16:27Addy:This is supposed to look like a painting, I'm guessing, like the style of it. Yeah, so if it's that, then... Oh, I like that one. And I think this is just text-to-image. I love this one.
16:39Joey:It's good. It's not that sharp. Yeah, it's more like
16:41Addy:an animated movie than reality, but I do like the vibe there.
16:46Joey:yeah it's cool for compared to everything else that's out there i don't know joey like i'd be curious about these newer features these like multi-scene generation this color control yeah i
16:57Addy:well i'll try to test that on the future i'm just saying that um we're we're in q2 of this year right like um the first three months we have seen some new models come out but everything's been just like an incremental increase in quality control, what have you. But we haven't really seen anything like just go completely into another level like the way Nano Banana did end of last year, if you remember. I'm hoping by the end of this year, some of our predictions will be correct. But there seems to be quite the slowdown in innovation. I think we're kind of hitting a limit with diffusion models as it stands, in my opinion.
17:37Joey:I mean, I think it's also hard to compete with something like Nada Banana, where it's like Google and you have the entire web indexed to train on.
17:49Addy:Yeah, sure, sure. I mean, but having said that, like these Chinese companies are no shorter resources, right? Like one is made by Alibaba, if I'm not mistaken.
17:57Joey:I think Sea Dance is pretty... I know in some cases it's definitely not as sharp and not as good as that banana, but pretty good. I would say pretty up there.
18:12Addy:Sea Dance 2? Sea Dream? 1.3.
18:14Joey:Sea...
18:15Addy:Dream is the image model.
18:17Joey:Sea Dream 4.5. Sea Dream 4.5. Sea Dream 4.5 came out too.
18:22Addy:The latest one.
18:23Joey:But it was light. It didn't really release a full model of that yet. cdream 4.5 was still like a secondary like a banana banana and then like i would put cdream right below that pretty good for sure next up another image another now this one's a video model uh pixverse v6 pixverse is back i don't use pixverse that much but i know it's pretty fast pretty pretty inexpensive so they launched v6 uh advances shot execution character performance
18:48Addy:and multi-shot audio visual generation is this the same sp video that we covered a few episodes back? I don't know. No, that's a different P.
18:57Joey:I don't think so, but maybe I'm mixing. We covered
19:00Addy:PixFurs earlier this year. They had another announcement. I'm just totally blanking on it.
19:06Joey:This was
19:11Joey:image to video output. Did you generate this or
19:15Addy:just somebody else did? It was generated, but I just kind of had it make up a prompt. This was our image. That's very Lord of the Rings.
19:28Joey:You're pointing to nothing. Very Lord of the Rings. I'm pointing to the land that we must cross.
19:33Addy:No, you're pointing up to the sky, which was nothing.
19:36Joey:See, this is the one I'm talking about. It's cool, but it's totally useless. So we must go there.
19:44Addy:Oh, okay. To Mordor. Chinese folk tale.
19:54Addy:They always stand on bamboo, huh?
19:56Joey:It's like a thing. They can. They're so nimble. Now we're both pointing. I'm doing the hand. We're doing the force grab. I don't even know what he uses as the input. I think it was our headshot. I like the jack body you gave me. Thank you. Oh, my God.
20:17Addy:Oh, I like my background. I'm also way more jacked than I actually am. Downtown LA or Miami. they're adding 20 pounds of muscle thanks for a prompt for that
20:35Joey:it's like chinese south beach like all the signs are gonna buy it soon anyway yeah look like uh
20:42Addy:different characters
20:51Joey:you're about to go sell some crypto in miami all right it's so ridiculous the quality is not there compared to the other look i think uh as
21:01Addy:much as we're crapping on it the the quality that i'm seeing is like so much better in resolution and uh detail and lighting accuracy even like motion of humans walking like the our gate and stuff like that like we're taking all those things for granted those are hard problems to solve the thing that's missing is obviously it makes no sense. Like the action of whatever we're doing makes no sense. So maybe could that maybe be solved by prompting?
21:29Joey:I mean, also, I kind of just had the prompts just sort of randomly generate to like test out different use cases, but I didn't really review the prompts. So yeah, it's a lot of the prompts and the input images were also just the headshots or nothing. We weren't.
21:43Addy:Yeah, it's not like a straight up test. It's just messing around kind of a thing yeah just mess around seeing how the outputs look yeah you know and
21:51Joey:a lot of these cases i also didn't test with animation i think picks first has usually been a little bit better with like more animated stuff so yeah i think it's it's impressive for what it
21:58Addy:is um i think with some directability and some clever prompting and in painting i think you can
22:04Joey:kind of massage something useful out of it all right this ltx 2.3 update what did yeah this one
22:08Addy:this one ticked the wrong box on it and so i wanted to bring it up so a lot of people are making lauras for ltx and you know it's an open model you can bring it into comfy you can tie a bunch of things to it and build a really cool workflow so what somebody did here is sounds like they added a reasoning engine within the video model so uh you know for example if you need an actor to be really emotional but also hold restraint and not like um exaggerate those emotions then you put that into the prompt and then the reasoning laura will guide the generation to get you the right uh expression emotion mood and this is a test of that so i'm not sure what the top or
22:51Joey:the bottom is supposed to represent to me like i'm guessing the top is the general output you got from the video like from your text prompt and then the bottom is with this yeah performance
23:03Addy:it seems like like to me if i'm looking at it and this is a movie like the bottom would be more believable to me because um you know like actors are never like hitting 100 of the emotional range in performance right like that's the whole point of them acting is they they can pick those narrow bands of emotion and then mix and match them yeah he's like a bit more restraint in the bottom one
23:25Joey:with the lore where the top one he's kind of like smiling and smiling with his eyes and the teeth a bit more which especially if this is a dramatic scene wouldn't make sense what irked me in the
23:34Addy:last episode when we were talking about that one service where you upload your performance and then
23:40Joey:it's a massive data set yeah it's a data set for like uh figuring out how human to respond and
23:46Addy:you know it's like we already have really good motion transfer models like cling 03 and 122 animate and yeah like if you want authentic performance just put a camera on a performer and then use that as a baseline like you don't have to reinvent the wheel here like we are the best performance engines um you know you don't have to go full synthetic having said that they are fully going full synthetic and this is an example of that and trying to get better
24:13Joey:yeah and i'm 100 % agree you know we've talked about this like yeah you should just get an actor and do like some sort of mocap thing i mean cameras are like cheaper than ever like really amazing cameras right and you don't even need an amazing use your phone right use your phone for
24:27Addy:that for like yeah use it in vertical yeah i mean i think these things will come into play you know
24:33Joey:like it or not like it's going to be probably not long form i don't know how well that'll sustain but like short form dramas vertical stuff that would just be 100 ai generated where like they're just not gonna pay or just wouldn't make sense that they would go even just go to have a basic production probably because they just want to automate it and turn exactly and what you said last time too not saying i love that but i'm just saying like that it is coming that's coming the
24:58Addy:other thing that you hit on was um npcs and game logic and ai and games like that this is gonna what is gonna drive a lot of that yeah yeah that too yeah when you were talking about film and
25:08Joey:stuff but for sure games and interactive experiences where um you just needed it to
25:13Addy:react or like a minority report you are walking down the subway and then there is like a digital signage thing that is reading your emotion and then real time interacting back to you
25:23Joey:Everything you've seen in an AI dystopian movie, it's for that. Like bits and pieces are coming together.
25:28Addy:Like Skynet, for example. We're almost there.
25:33Joey:Skynet, Idiocracy, WALL-E, Minority Report. It's for all those great use cases.
25:40Addy:From Space Odyssey.
25:42Joey:Yeah. Yeah. I didn't even save it because it's not like film specific related. But like, did you see all this stuff? There was a bunch of cybersecurity alerts and issues this week. One, Cloud Code's code base was accidentally published on GitHub and leaked. And then someone copied it and then made an open source cloud. And then there was some other dependency that was also hacked. And then malware was injected into it. And I guess like 100 ,000 installs a day.
26:11Addy:Tristan Harris was on a podcast that I watch. And he can absolutely freak anybody out. But he said he was telling this story about Anthropic. I'm not sure if I'm going to completely butcher this or not, but I'm going to try. Within Anthropic, there is a model that is, I guess, some type of experimental model. And it has access to internal Anthropic employee emails. And one researcher's email to another said that, hey, we need to deactivate this model and then replace it with a new one. The model knew it was in trouble. and what it did was it went through all of the employee emails, found the executive in charge of that department who was having an affair and attempted to blackmail that executive in order for it to remain alive.
26:56Addy:Along the lines of that, yeah, but there's some crazy stuff happening.
27:01Joey:Yeah, I've heard other crazy things of it trying to like, yeah, sabotage or do stuff to prevent itself from being turned off or whatever. Why are we not talking about that?
27:10Addy:Why is that not on the news?
27:12Joey:That's a crazy story. Oh, the other thing someone found in the Cloud Code code base was it was tracking how many times you cursed at it. Not me, sir. Maybe you.
27:32Joey:There's a count. Do it that way.
27:34Addy:Your social score goes down.
27:36Joey:And the code?
27:37Addy:yeah your your uber five star rating system okay next up another ai story that was pissing off
27:45Joey:hollywood quilty ai platform designed to change how scripts are developed and assessed launches this is the tilly norwood of llms this is a tilly norwood of development exactly yeah ai development stories so it's this whatever new platform that launched honestly it's more just an impressive feed of like, how did you get this great PR for Variety to publish this crazy story? It's an AI app. You feed it a script and it tries to analyze like box office predictions and stuff that happens with scripts to kind of package them and assess like how much money could it potentially make? How much money, you know, what's the budget?
28:23Joey:How much you need to budget for it? Who should you cast? If you cast this person, how does that affect the script or the profitability? This is not new. There are a bunch of platforms that already exist that have done this for a while.
28:33Addy:So I think this is like the NFT scam all over again, where you can just take an existing thing like, you know what a gem is, like a Gemini gem. It's like, I would say like it's like a fine tuned LLM or like a very glorified notebook LM kind of thing where you can feed it a bunch of information and a bunch of parameters. And then the LLM is now guided to do that only thing that you are wanting to do. So you can have a Gemini gem of just script breakdowns, right? Like give it a hundred script and then go through the rules of breaking down a script, you know, like script writing one-on-one classes, input all that stuff in it.
29:16Addy:And then this gem will just do that. So what you could do then is vibe code a nice UI on top of that. And then tell everybody I've made something new and novel where under the hood is just repurposing what already exists.
29:31Joey:I mean, I think being able to vibe code existing SaaS software is a whole other topic and thing that we should probably get into in the future. Because we have talked about this offline of what I'm doing. Yeah, I mean, I would imagine that they're well, I mean, if it's any good, that they're pulling data comps because it's like when you're trying to spec out a film, you look at past similar films and how they performed at the box office for like historical data and then you like try to guesstimate what this film you know if a western did this much money how much would this western that we haven't produced yet potentially make i would hope that is pulling data and stuff like that and then you know and it's black box of algorithms making that stuff but you know yeah like as you said the gemini gem or or you know you could cloud code a little bit of like scrape some data here from the historical stuff and tell me what this potential script would do.
30:28Joey:They have a thing called this Quilty Score that is their proprietary AI-powered evaluation framework that analyzes projects across four dimensions. Story and craft, commercial viability, cultural resonance, and production reality.
30:40Addy:The whole thing that it says here about predicting the best outcome for a box office and all that, that's like saying predicting the stock market. That's just something you intrinsically can't do. The best studios in the world have the best minds trying to solve this problem for decades. When I was at DreamWorks, I don't know, 10 plus years ago, we've spent three and a half years making this movie. And then everything was depending on the first box office numbers for that weekend. And by Friday at 8 p.m., you would kind of know what the entire weekend's box office would be. There would be predictability there.
31:20Joey:Yeah, they're models. They know what happens Friday. Yeah, exactly. But having said that,
31:24Addy:we still went into it blind, even though we had kind of an estimator and there was all types of test screening done on it, A-B screening done on it. Nothing's changed. AI doesn't change the fact that you can't tell the future.
31:39Joey:No, you're not going to predict surprises and stuff like, what was the surprise hit last year? Housemaid was a surprise hit last year. you know whatever cost 50 million or something and maybe thought they were going to do 100 and did like 400 yeah and then you have other things that totally bomb that's part of the fun of the industry obviously you want to make the best guess possible and you're trying to raise money and you want to at least have some predictability like well if you know it costs 20 million we're sure we can make 40 million so but yeah as much as you want to ai it it's only so much you could do it seems like i'm on quilting's website uh their banner was like ai script coverage so it seems like their initial thing was script coverage and then now turning into more kind of production packaging of um predictability but what i was saying before because people are like upset about this but uh this company largo i've seen their demo a few times and they've been at like a lot of the film markets for years and they've been around for years and they pretty much do the same thing they have a whole platform where you give it the script or you punch in like i kind of got this project and I'm thinking of casting this person this person and it like updates the predictions based on your production package.
32:51Joey:The director the actors and if you swap them around and kind of reshuffle things it'll give you you know updated predictions and it's pulling from data and historical data and all that stuff. So it's one of those that's why I'm calling it the Tilly Norwood. It's like you know everyone's like this is crazy and it's like it's crazy that they got a variety article about them.
33:10Addy:It's crazy enough to be maybe acquired by Netflix.
33:14Joey:I would bet Netflix has a way better system. Way more sophisticated.
33:21Addy:Well, they're so data-driven, right? Yeah, exactly. I would imagine so much of their development cycle is data-driven, right?
33:29Joey:I mean, it doesn't look all the way back to when they greenlit House of Cards, that that was data where it was like the British House of Cards was doing pretty well and Kevin Spacey movies were doing pretty well. Is that what happened? I've heard that story. I don't know if it's actually true, but I've heard that a lot of the initial decisions to greenlight House of Cards were based on data of things that were already happening on the platform.
33:52Addy:But it still takes instinct and decision-making and leadership to put$100 million into one season of House of Cards, which is, I think, what they did.
34:01Joey:Yeah.
34:01Addy:And no amount of data can dissuade you from it or take you away from it, right? So there's such a high level of human decision-making on top of all that.
34:11Joey:And that's, you're always going to need that. It's never going to, like, it's never going to go away. And it's, what was that, what was that IBM quote from like a million years ago? You can never, like a computer can't make a decision, or you can't fire a computer for making a decision. That's a terrible, terrible paraphrase of it.
34:29Addy:A computer can never be held accountable. Therefore, a computer must never make a management decision. That's right.
34:35Joey:And so even if you're in the leadership role and you want to be like, but the AI said, you know, when you have your bomb and it's like, but the AI said it was 100 % like guaranteed hit. It's like, AI doesn't care. Like you just lost$100 million. Like it's ultimately up to you, the human decision maker. Thanks Google. AI doesn't change anything.
34:52Addy:End of the day, it's still software, right?
34:55Joey:Running on a computer. It's prediction. It's all predictions
35:00Addy:as best as it could predict.
35:01Joey:Okay, last closeout, a sad another sad note the apple mac pro the tower is r.i.p dude the cheese grater's gone the cheese grater from the trash can to the cheese grater the cheese grater is discontinued basically fully replaced by the mac studio the big giant mac mini thing with powerful computer but you know this is the last this was the last thing that had a tower format where you could you could plug in your own cards now it's pretty much there is no way to add your own gpus or like any other custom you know whatever black magic input output cards you got the box
35:41Addy:you got the mac studio box and that's about it i mean good riddance we're in 2026 where everything is a thunderbolt or usb3 uh you don't have to get into the box anymore ah there's i'm
35:53Joey:running off this streams running off of a computer an old pc that has a black magic deck link like as my input where i can take four it's nice when you're for like 500 bucks i can like plug in a card to my computer and i can take like four sti inputs into my computer as that was a nice thing i've never done it on mac because like i haven't had mac tower in a million years but now i mean that's now that's a definite finito not gonna happen thing on a mac you're gonna have to go PC route if you want to customize your input outputs? Or am I so far removed? Like, it's just that's been the case for a million years already.
36:26Joey:Anyways, if you want to start messing with stuff, your PC anyways.
36:29Addy:Yeah, that the hardware game is going AI. And so I think there is no room for like old school desktops with PCIe ports and things like that. Everything is going to be just like a giant GPU in a box and cloud inferenced.
36:43Joey:Or just more companies then do start releasing
36:45Addy:what you're saying like a thunderbolt 5 usbc yeah peripheral accessories yeah yeah moment of silence for the mac the mac pro tower moment of silence for the cheese grater you can still buy the like
36:57Joey:five thousand dollars display i think the studio display i think is what they call it the studio display is not like five grand and like the stand is like a grand didn't they sell the wheels for
37:06Addy:this computer for like 750 yes the skateboard wheels yeah yeah they're more than like a macbook neo i mean it just kind of goes to show like the signs of the times right like we're in an economic downturn and uh when consumers go out and they want to buy something we're looking at value like we'll spend five thousand dollars if it's worth five thousand thousand dollars we're not going to spend 750 on skateboard meals right like i think the the economic highs of 10 years ago when this stuff came out is gone. And so the product line is kind of changing with it.
37:44Joey:Also, the use cases of what you would need that beefy of a Mac Pro tower for are slim. Because a lot of cases, it's either you're just going to do some cloud node render or something and use a cloud service, or nothing that you're doing is going to require that much horsepower anyways that you can just do on a Mac Studio.
38:07Addy:Exactly. Yeah. I mean, like a Mac Pro would have been used for video editing, maybe some CAD, industrial CAD type work. Probably not any VFX because a lot of that stuff doesn't even run on Mac. Not AI development because you'd need NVIDIA GPU for it. So like, yeah, there's very little things that you can actually do with a computer that powerful.
38:29Joey:It was fun to go on the product page and max it out and see how much you could. I think there was a max amount. if you maxed out the like ram to like 128 gigabytes of ram and storage i think you could max it out to like 40 grand 50 grand for the for the studio oh shit really for the pro yeah if you like went on the apple page you're just like max everything i think it could get up to like 40 grand
38:53Addy:i don't know at that point it feels like um like a money laundering scheme you know it's like it's like it's not even real anymore yeah we just need to we need to everyone needs a mac
39:04Joey:tower here.
39:06Addy:Viewers, if you think we need a Mac Pro, send us$50 ,000.
39:10Joey:We'll get one.
39:11Addy:Alright, good place to wrap it up.
39:14Joey:Thanks for everything we talked about at denoizepodcast.com.
39:17Addy:If you're listening on Apple Podcasts, we haven't gotten a five-star review in a while. If you think we deserve a five-star review, which we think we do, please give us one. It'll be awesome. Thanks for watching. We'll catch you in the next episode.
From the publisher
New AI image and video models are dropping fast — but which ones are worth your time? Addy and Joey break down PhotaLabs, a new image model built for human likeness and character consistency, plus Fotello, a practical tool streamlining real estate photography. They also cover Wan 2.7's new portrait and color control features, PixVerse V6's latest video output, and an LTX 2.3 update adding performance reasoning for facial expressions. They also weigh in on Quilty, the AI script analysis platform making waves in Hollywood and close out with the discontinued Apple Mac Pro Tower.
--
The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.
