Rendered Real: The Noir Starr Podcast
"Rendered Real: The Noir Starr Podcast" dives into the intersection of high fashion, artificial intelligence, and authentic representation. Hosted by the visionary team behind Noir Starr Models, each episode explores how the digital modeling revolution is reshaping beauty standards, brand storytelling, and the future of talent.
Rendered Real: The Noir Starr Podcast
Episode 47: The Rise of AI Supermodels: Engineering Digital Influence
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Episode 47: The Rise of AI Supermodels: Engineering Digital Influence
In this episode, we explore the high-stakes evolution of the modeling industry. We are witnessing the birth of the AI Supermodel—digital avatars that aren't just pixels, but precisely engineered engines of global influence.
As we hit mid-2026, the logistical shift is undeniable. Brands are moving toward synthetic talent that offers 24/7 availability, zero travel costs, and "locked-in" visual consistency across every frame of a campaign. But as these perfect renders dominate our feeds, we have to ask: what is the cost of removing the human element from the face of fashion?
You know, um it's funny how quickly an illusion can just completely shatter.
SPEAKER_00Oh, absolutely. It only takes one tiny mistake.
SPEAKER_01Right. Like think about watching a massive big budget blockbuster movie. When the CGI is bad, it's literally all you can focus on, you know?
SPEAKER_00Yeah, you stop paying attention to the plot entirely.
SPEAKER_01Exactly. The physics are just, I don't know, slightly off, or the lighting doesn't quite match the background, and suddenly you're yanked right out of the story. The magic is just gone. But when the CGI is genuinely good.
SPEAKER_00You don't even know it's there.
SPEAKER_01Yeah. You just accept the dragons or the spaceships as reality. You don't even question it.
SPEAKER_00Aaron Powell Well, the absolute best special effects are the ones that uh they never actually register as special effects. They just seamlessly weave themselves into the fabric of the world you're looking at, and they completely bypass your brain's natural skepticism.
SPEAKER_01And that razor-thin line between a broken illusion and seamless reality is, well, it's the exact tension we're looking at today. Because the$312 billion luxury fashion industry is currently in the middle of a very public, incredibly expensive identity crisis over artificial intelligence.
SPEAKER_00It is a fascinating collision of art and engineering right now. And uh I think the perfect starting point to understand this panic is the absolute disaster of Balenciaga's fall 2025 campaign.
SPEAKER_01Oh, wow. Yeah. That was a spectacular failure.
SPEAKER_00Truly historic.
SPEAKER_01I mean, Balenciaga decided to go all in, right? They used entirely AI-generated models for a major global campaign. And the backlash wasn't just immediate, it was visceral.
SPEAKER_00People were so angry, fashion critics were practically tripping over themselves to slam the imagery. I mean, the word that kept coming up was soulless.
SPEAKER_01Soulless, exactly. It was such a colossal misstep that within 48 hours, which by the way is barely a fraction of a second in corporate fashion time, this massive luxury giant panicked, walked the entire campaign back, and spent an absolute fortune to completely reshoot everything using real human models.
SPEAKER_00It was a very, very public retreat. But you know, to really grasp the complexity of the situation, we need to contrast Balenciaga's catastrophe with some internal research coming out of ASOS, the massive online retailer.
SPEAKER_01Right, because they had a totally different experience.
SPEAKER_00Exactly. While Balenciaga was dealing with their PR nightmare, ASOS was quietly running these blind tests. They were putting AI models up against traditional human photography. And the results from those tests are, frankly, staggering.
SPEAKER_01I read this part of the report and my jaw dropped.
SPEAKER_00Right. 67% of Gen Z consumers couldn't even tell the difference.
SPEAKER_01Over two-thirds. Over two-thirds of the demographic that practically lives their entire lives online couldn't spot the fake. That is wild.
SPEAKER_00It really puts the Balenciaga backlash into perspective. It does.
SPEAKER_01So that's our mission for this deep dive. We've got a massive stack of industry analyses, ethical codes of practice, and marketing case studies to really understand luxury fashion's AI dilemma.
SPEAKER_00And it is quite a dilemma.
SPEAKER_01It is. Because this isn't just a story about cool new tech. We're going to explore the brutal economics driving this massive shift, the incredibly manipulative psychological tricks being deployed by virtual influencers, and uh the precarious ethical tightrope that brands are walking right now.
SPEAKER_00Aaron Powell Yeah. And to understand why any brand would risk the kind of backlash Balenciaga faced, we really first have to look at the undeniable leap in the technology itself, and of course the sheer crushing weight of the economics behind it.
SPEAKER_01Okay, let's unpack this. Because we aren't talking about the glitchy, you know, weird-fingered AI from two or three years ago.
SPEAKER_00Aaron Powell No, not at all. That era is ancient history at this point.
SPEAKER_01Right. The sources highlight that we've crossed a massive threshold with things called diffusion models. We're looking at platforms like stable diffusion XL, proprietary systems from Deep Agency, and uh the new black. But I need you to explain how this actually works because it's not just pasting pixels together from a database, is it?
SPEAKER_00Aaron Powell No, it's it's a fundamentally different process. Diffusion models learn in a really counterintuitive way. So imagine taking a crisp, high-resolution photo of a runway model and slowly adding digital noise to it.
SPEAKER_01Like like TV static.
SPEAKER_00Exactly, like the static on an old television.
SPEAKER_01Yeah.
SPEAKER_00You add it until the original image is completely destroyed. Now the AI studies that specific process and then it learns how to reverse it.
SPEAKER_01Wow.
SPEAKER_00Yeah. So when you ask it for an image, it doesn't pull from a database of photos or just stitch things together. It starts with a canvas of pure static and literally hallucinates a photorealistic human being and a high fashion coat out of the noise.
SPEAKER_01That is just that is mind-bending.
SPEAKER_00And that's how we're getting images generated in over 8K resolution now.
SPEAKER_018K resolution. But the technical detail in these reports that really caught my eye, the thing I couldn't get over, was this term, subsurface scattering.
SPEAKER_00Ah, yes. Subsurface scattering is the absolute holy grail of digital rendering.
SPEAKER_01What does that actually mean though?
SPEAKER_00To put it simply, when light hits human skin in the real world, it doesn't just bounce off the surface like a mirror, right? It actually penetrates the outer layer, scatters around in the dermis underneath, interacts with your blood vessels, and then reflects back out.
SPEAKER_01Okay, so it's absorbing and reflecting at the same time.
SPEAKER_00Exactly. That complex interaction of light is what gives human skin its glow, its depth, and you know, its vitality.
SPEAKER_01And these AI platforms are now mathematically calculating that exact physics of light hitting a human dermis.
SPEAKER_00Down to the microscopic level. And importantly for the fashion industry, it applies to fabric physics as well.
SPEAKER_01Right, because clothes matter just as much as the models.
SPEAKER_00Precisely. The algorithm understands exactly how heavy a silk drape should fall based on gravity. It knows how light should catch the sheen of genuine leather versus synthetic leather, or or even the microscopic texture variations in a cashmere sweater.
SPEAKER_01That is wild.
SPEAKER_00It even understands how these materials respond to a camera flash dynamically.
SPEAKER_01Which brings us to the brutal math of it all. Because the financial numbers in these reports are frankly staggering. If you're doing a traditional fashion shoot, you're just bleeding money.
SPEAKER_00Oh, the overhead is massive.
SPEAKER_01It is. You have to rent a studio, you pay for casting, the models, the photographers, lighting technicians, caterers, hair and makeup artists, and then you have weeks of expensive post-production.
SPEAKER_00It adds up incredibly fast.
SPEAKER_01Yeah. The sources estimate a single traditional campaign costs anywhere between$50,000 to$200,000.
SPEAKER_00Aaron Powell And that's just a massive logistical and financial undertaking for ultimately just a handful of usable images.
SPEAKER_01Trevor Burrus Right. But then you look at the AI Visual Commerce platform as mentioned in the reports, there's one called Rewarks, their starting price,$9.90 for the first month.
SPEAKER_00$10.
SPEAKER_01$10.
SPEAKER_00The contrast is almost absurd when you put those numbers side by side.
SPEAKER_01It's completely absurd. I mean, a mid-sized fashion brand that shoots, say, 500 seasonal items, they could save$2.3 million every single year just by switching to AI-generated imagery.
SPEAKER_00And beyond the money, there's the speed factor.
SPEAKER_01Yes. Zara's parent company, Inditex, they used AI to cut their time to market by 60%. They went from refreshing their catalogs quarterly to refreshing them monthly. And McKinsey is out here estimating that generative AI could add between$150 and$175 billion in value to the retail sector by 2030.
SPEAKER_00Which is why there is such an overwhelming pressure from boards and investors. They see the bottom line, they see that speed, and they demand adoption.
SPEAKER_01So I have to push back here. Because if I'm a brand executive and I can get 8K resolution, perfect subsurface skin scattering, and flawless cashmere physics for$10 a month instead of$200,000, why wouldn't I literally fire every single photographer on my payroll tomorrow morning?
SPEAKER_00Well, because, as the industry analysis beautifully points out, good enough is not luxury's operating principle.
SPEAKER_01Okay, explain that.
SPEAKER_00Sure. The mass market is built entirely on efficiency, right? Fast fashion needs speed and cost cuts. But the luxury market operates on an entirely different psychological frequency. It relies on what the industry calls narrative authenticity.
SPEAKER_01Narrative authenticity. Break that down for me.
SPEAKER_00Think about the transaction itself. When you pay$5,000 for a coat, you aren't just buying wool, thread, and buttons.
SPEAKER_01Right. You're buying the brand.
SPEAKER_00You're buying a story. You're buying the mythos of human craftsmanship. The idea of an artisan in a European atelier spending hours perfecting a single stitch. In fact, Gucci's creative director recently noted that their campaigns require human vulnerability.
SPEAKER_01Human vulnerability, I like that phrase.
SPEAKER_00It's crucial. A computer code, no matter how flawlessly it renders the drape of that coat, it cannot fabricate true human vulnerability. The moment an AI model wears that$5,000 coat, it breaks the spell.
SPEAKER_01Ah.
SPEAKER_00And reminds the buyer that this is a highly scalable, digitally manufactured commodity, not a rare human-crafted artifact.
SPEAKER_01That makes total sense. The illusion relies on the exclusivity of human effort. But okay, if human vulnerability is the ultimate luxury currency, how do we explain a massive glaring contradiction in the market?
SPEAKER_00You mean the virtual influencers?
SPEAKER_01Yes. Because on one hand, consumers supposedly demand this human touch, right? But on the other hand, those same consumers are out there building incredibly deep emotional connections with entirely fabricated AI personalities online.
SPEAKER_00Aaron Powell It's the virtual influencer paradox. It completely flips the narrative authenticity argument on its head.
SPEAKER_01It really does. Take Lil McKella. We all know the 19-year-old half-Brazilian, half-Spanish model who lives in LA, except obviously she doesn't exist. She's an AI avatar.
SPEAKER_00Created by Brood, which is a transmedia studio that recently hit a$125 million valuation entirely based on her success.
SPEAKER_01Which just blows my mind. Michaela herself has an estimated net worth of$10 million. Time magazine named her one of the 25 most influential people on the internet, right alongside actual flesh and blood human beings like Rihanna. It's what? And she charges an estimated$10,000 for a single sponsored Instagram post, who is paying$10,000 for a rendering.
SPEAKER_00Honestly. Everyone from Prada to Sansun.
SPEAKER_01Wow.
SPEAKER_00And the ecosystem is expanding rapidly beyond just her. You have Ludo Magalu, who completely dominates retail influence in Brazil. You've got Emma, who's famous in Japan for her signature pink bob and hyper-realistic luxury shoots.
SPEAKER_01Oh, yeah, I've seen her.
SPEAKER_00There's Nunari, who actually advocates for social and environmental causes, and Shudu Graham, who is widely considered the world's first digital supermodel.
SPEAKER_01The psychology behind why they work is honestly slightly terrifying. The Noir Star Models report talks about how these avatars are engineered for manufactured relatability. But how does that actually work in practice, like mechanically?
SPEAKER_00The mechanism is what's truly fascinating. When a human influencer shares a personal struggle online, it's usually a genuine flaw or just, you know, a uniquely bad day.
SPEAKER_01Right.
SPEAKER_00But when a virtual influencer shares a flaw, it's a result of a massive data pipeline.
SPEAKER_01Wait, really?
SPEAKER_00Oh, absolutely. The algorithm constantly scrapes social media platforms for sentiment analysis. Let's say there's a huge spike in Twitter data around imposter syndrome or burnout among Gen Z.
SPEAKER_01Okay.
SPEAKER_00That data is fed into generative models to script an Instagram caption for Lil McKella complaining about feeling overwhelmed. Her vulnerabilities are algorithmically generated based on real-time data sets to perfectly mirror human insecurities back at the audience.
SPEAKER_01So it's literally weaponized empathy. And for the brands hiring them, I mean they're the ultimate employee. We talked about the scalability of the tech earlier, but the thing about the logistics. One virtual model can headline a product launch in Tokyo at dawn, walk a digital runway in Paris at noon, and star in a New York ad campaign in the evening. Trevor Burrus, Jr.
SPEAKER_00Without ever getting jet lagged.
SPEAKER_01Exactly. They never age, they don't need craft services, and crucially, they never go off script and cause an unpredictable PR scandal.
SPEAKER_00Aaron Powell Right. They deliver all the edgy, modern cultural cachet of a young influencer with absolutely zero of the corporate risk.
SPEAKER_01Here's where it gets really interesting. This is the hypocrisy I just can't wrap my head around. We aggressively reject an AI model wearing a Gucci coat because it breaks the spell of human craftsmanship. But then millions of us eagerly follow a fake 19-year-old who stages fake dramas.
SPEAKER_00Oh, the hacking stunt.
SPEAKER_01Yes. When Brood literally orchestrated a fake Instagram hacking by another CGI character named Bermuda just to drive engagement, why do we punish Balenciaga for using AI, but reward Brood with a$125 million valuation for doing the exact same thing?
SPEAKER_00Well, it comes down to the fundamental difference between static product replacement and transmedia storytelling.
SPEAKER_01Okay, unpack that.
SPEAKER_00When Balenciaga used AI, they were using it to quietly replace human models in a heritage context. They were trying to pass off the artificial as real to save a buck while still charging a premium for the human touch of the clothing. The consumer feels deceived by that.
SPEAKER_01Because it's a bait and switch.
SPEAKER_00Exactly. But with Lil McKella, the AI isn't hiding. The AI is the story. She succeeds wildly because she leans into her own digital narrative. People don't follow her thinking she's flesh and blood. They follow her the same way they follow a character in a television show.
SPEAKER_01Oh, I see.
SPEAKER_00It's interactive entertainment. It's transmedia storytelling that spans across Instagram, YouTube, and real-world brand partnerships, all openly acknowledging the fiction.
SPEAKER_01Ah, so the transparency of the fiction makes it acceptable, which actually perfectly explains how brands are trying to implement this tech safely today. Right. If pure deceptive AI models are too risky for legacy luxury, but those$10 a month economics are too good to ignore, they have to find a middle ground.
SPEAKER_00And we're seeing the industry fracture into very specific implementation tiers based on brand prestige to navigate that middle ground.
SPEAKER_01Let's walk through this hybrid tier strategy because it's the actual roadmap brands are using right now. At the very top, you have hyper-luxury brands like Chanel, Hermes, Dior. According to the analyses, they are using AI strictly behind closed doors.
SPEAKER_00Yes. They use it for internal mood boarding and technical design iteration, but never in public-facing campaigns. Diluting that heritage mythos is simply too massive a risk for them.
SPEAKER_01Then you step down a bit to accessible luxury brands like Michael Kors or Coach. They're using a hybrid model. They'll use AI to generate the thousands of standard e-commerce catalog shots for their website, but they still shell out for human supermodels and photographers for the big glossy hero campaigns on billboards.
SPEAKER_00It's a pragmatic split.
SPEAKER_01And then at the mass market level, brands like ASOS are going fully AI first for speed and scale.
SPEAKER_00The most innovative approach, though, the one that really threads the needle, is blending the two in front of the consumer through digital twins.
SPEAKER_01Aaron Powell Right. The HM 2025 campaign is the ultimate example of this. They ran a massive spring campaign where they photographed 30 real human models, but then they used AI to create hyper-realistic digital twins of those exact models.
SPEAKER_00Aaron Powell And they put the human and the AI version side by side in the ads.
SPEAKER_01Aaron Powell Which is wild. But the crucial detail here is that HM paid the human models for the use of their likeness. They allowed the models to retain their image rights, and they clearly openly labeled the AI content for the consumer.
SPEAKER_00Trevor Burrus That upfront transparency is the only reason it worked instead of causing a Balenciaga style backlash.
SPEAKER_01And this digital twinning concept goes way beyond just static photos. It's getting into hyper-personalization. Like Nike did this incredible campaign for their 50th anniversary.
SPEAKER_00Oh, the Serena Williams one.
SPEAKER_01Yes. They used AI to analyze Serena Williams' playing style from decades ago, mapping her old swings and footwork. Then they had the current Serena Williams play a virtual tennis match against a digital twin of her past self.
SPEAKER_00It's brilliant because it uses the technology to visualize a narrative about discipline and evolution that literally could not be told through traditional filmmaking.
SPEAKER_01Exactly. And Cadbury did something similar in India for the WALLI. They used AI facial mapping of the massive Bollywood star Shah Ru Khan. But instead of just broadcasting a generic national commercial, they allowed local corner stores to generate customized ads where a digital Sharu Khan specifically promoted their small shop.
SPEAKER_00That hyper-local approach is so smart.
SPEAKER_01It drove a 35% lift in sales. And then you have B2B companies like Synthesia leaning all the way in, making the AI Avatar tool the central focus of their marketing, proving that the tech itself is the product.
SPEAKER_00Yeah.
SPEAKER_01Even Red Lobster got in on it, generating quirky AI country songs about their biscuits, dubbing it Cheddar Bay Ai.
SPEAKER_00Well, uh Cheddar Bay is certainly a creative choice, but it preves the overarching point. When brands are playful and entirely transparent about the AI, consumers will happily engage with it.
SPEAKER_01I hear that, but I have to challenge this middle ground for a second. These campaigns are clever, sure. But doesn't this hyper-personalized digital twin stuff feel incredibly dystopian?
SPEAKER_00In what way?
SPEAKER_01Well, if Cadbury is using a digital Sharu Khan to talk directly and specifically to my local neighborhood corner store, aren't we blurring the line of reality so much that consumer trust will eventually just snap?
SPEAKER_00That raises a really critical point. And it connects directly back to what psychologists call the uncanny valley.
SPEAKER_01Right, that creepy feeling.
SPEAKER_00Exactly. As these digital avatars get more realistic, and as the hyper-personalization gets more granular and targeted, human beings naturally feel a primal sense of unease. Our brains subconsciously detect that something in the interaction isn't quite right.
SPEAKER_01So how do brands avoid that?
SPEAKER_00To avoid that complete snap in consumer trust, success requires total unwavering transparency. The exact moment a brand tries to pass off an AI as an authentic human interaction, when they shift from entertaining the consumer to tricking them, they risk a permanent loss of trust.
SPEAKER_01And that looming threat to consumer trust combined with the very real immediate world harm to human professionals is exactly why the industry is scrambling to write the rules before the entire system breaks.
SPEAKER_00The human cost here isn't a theoretical debate for five years down the line. It's happening today.
SPEAKER_01The Noir Star Models report highlights this dark side incredibly well. It's not just the multimillion dollar supermodels whose jobs are at risk. It's the working class professionals of the fashion world.
SPEAKER_00The people behind the camera.
SPEAKER_01Yeah. Real photographers, lighting techs, makeup artists, and stylists are losing out on vital contracts to code that costs 10 bucks a month. And beyond the massive economic displacement, there's a heavy psychological toll happening with beauty standards.
SPEAKER_00Algorithmic beauty standards are uniquely insidious because of how they're built.
SPEAKER_01Because these AI models aren't based on human genetics, they're optimized purely for engagement. They aggregate millions of data points to mathematically homogenize human features into impossible ideals of beauty.
SPEAKER_00Which is so damaging.
SPEAKER_01It is heavily distorting self-perception, especially in younger audiences who are scrolling through feeds, not realizing they are comparing their real bodies to mathematical equations.
SPEAKER_00And when you rely on algorithms instead of human judgment, you also lose cultural sensitivity and empathy.
SPEAKER_01Case in point, Calvin Klein. They ran a campaign featuring real-world supermodel Bella Hadeen alongside the virtual influencer Lil McKella. The video depicted them kissing.
SPEAKER_00Right? I remember this.
SPEAKER_01Calvin Klein clearly thought they were being edgy and futuristic. Instead, they faced an intense, massive backlash from the LGBTQ community who accused the brand of queer baiting. It demonstrated perfectly how an algorithmically driven marketing approach, lacking actual human empathy and oversight, can carelessly exploit sensitive social issues for a quick viral profit.
SPEAKER_00Which is exactly why the industry is attempting to self-regulate before global governments step in and drop the hammer on them.
SPEAKER_01Enter the BFMA, the British Fashion Model Agents Association. They've drafted a comprehensive code of practice specifically for AI in the modeling industry.
SPEAKER_00The guardrails they're proposing are actually quite rigorous.
SPEAKER_01The five key pillars they lay out are fascinating. First, absolutely no misleading audiences with undisclosed digital replicas. If it's AI, it must be labeled clearly.
SPEAKER_00Which addresses the transparency issue perfectly.
SPEAKER_01Second, AI must not introduce bias into casting or representation.
SPEAKER_00Crucial for fighting those homogenized beauty standards.
SPEAKER_01Third, and this is massive for the talent, individuals must retain total ownership of their biometric data. A brand cannot scan your face and own your digital twin forever. That's a huge protection. Fourth, any data collected on a traditional chute cannot be secretly siphoned off to train a brand's AI data sets without explicit consent and compensation. And finally, a really interesting point. They demand that brands actually calculate and consider the massive environmental impact and server energy required for these huge generative AI workflows.
SPEAKER_00It's a strong framework. It really attempts to balance the inevitable march of innovation with basic human rights and fair compensation.
SPEAKER_01So what does this all mean? Because I look at this voluntary code, and my analogy is that it feels like the Wild West out there. And the BFMA is just a handful of deputies trying to build the very first sheriff's office.
SPEAKER_00That's a fair comparison.
SPEAKER_01Realistically, is a voluntary code of practice actually going to stop a massive fast-fashioned brand operating overseas from secretly training their AI on unconsenting models just to save$2 million a year?
SPEAKER_00It's a sobering reality check. You're looking at a profound tension between ethical guidelines and ruthless market efficiency. Legislation moves at the agonizing speed of bureaucracy, while AI evolves at the exponential speed of computation.
SPEAKER_01The tech is just too fast.
SPEAKER_00Exactly. Right now, the financial incentives to cut corners and replace human labor with a$10 subscription are vastly outpacing the speed of regulation. The BFMA is drawing a line in the sand, but it will take immense consumer pressure to force the broader industry to respect it.
SPEAKER_01And that's really the crux of this entire deep dive. We started by looking at the brutal economics, how subserface skin scattering and flawless fabric physics can save a brand millions of dollars overnight.
SPEAKER_00The math is undeniable.
SPEAKER_01We explored the engineered, manufactured relatability of virtual influencers who make fortunes by algorithmically pretending to have human flaws. And we ended on the desperate, urgent attempt to establish ethical guardrails before human talent is entirely displaced by server farms.
SPEAKER_00It's a massive paradigm shift in how we perceive authenticity, art, and commerce.
SPEAKER_01So the next time you're scrolling through your feed and you feel a pang of jealousy over a flawless model and a luxury campaign, or you feel deeply understood by an influencer's incredibly vulnerable, relatable post. Well, take a second look.
SPEAKER_00Definitely zoom in on the details.
SPEAKER_01You might just be forming a parasocial relationship with a server farm that has been perfectly mathematically optimized for your specific demographic.
SPEAKER_00Always question what you're looking at.
SPEAKER_01But I want to leave you with one final provocative thought to mull over. Throughout this entire deep dive, we've talked about brands carefully controlling these AI avatars to avoid PR disasters. But generative AI storytelling is becoming more and more advanced every single day. It is. What happens when these virtual influencers become fully autonomous?
SPEAKER_00Oh.
SPEAKER_01What happens to the cultural landscape when a synthetic celebrity goes rogue, not because of a staged PR stunt written by a human marketing team, but because the algorithm independently determines that creating a real-world controversy or expressing a radical opinion is mathematically the most efficient way to boost its own engagement metrics.
SPEAKER_00The implications of an algorithm deciding that outrage is the most efficient marketing strategy are staggering.
SPEAKER_01When the algorithm writes the script, we won't just be in the uncanny valley of how things look, we'll be in the uncanny valley of how things behave. Until then, keep looking closely at the screen, because the illusion is only going to get harder to spot.