
The AI Bodybuilder with the Hidden Eyes Who Got 48.7M Views, So I Built My Own AI Character
The AI character Derek Mercer is a fake "bodybuilder" on TikTok who never shows his eyes, and his first 12 videos combined got around 48.7 million views. This post summarizes and rewrites in Vietnamese the analysis of a creator who dissected that phenomenon and then built their own AI character. According to GHOST (@ghosstty_, 16/10/2026), there are four main points: why that character is so memorable, the character design process from reference sheet to video, costs around $155 for 10 videos, and why this wave could cool down fast. First is Derek's story, next is the process, finally is the cost and risk.
This post is based on the article "$48K and 48.7M Views From an AI Bodybuilder Whose Eyes You Never See. So I Built My Own AI Character" by GHOST (@ghosstty_) on X (16/10/2026), rewritten in Vietnamese by OPCAIOS, condensed and supplemented with notes for content creators in Vietnam. The numbers about the account mentioned here are counted or cited by the original author themselves; OPCAIOS has not independently verified them.
Who is Derek Mercer and where did the 48.7 million views come from?
Derek Mercer is a bodybuilder character who has never stepped foot in a gym: every video is AI-generated, and some captions even say straight out that it's made with Higgsfield, an AI image and video generation platform. According to GHOST (2026), the TikTok account has around 170 thousand followers and 4.2 million likes, follows nobody, and the person behind it has never shown their face, name, or voice.
Before taking the numbers at face value, the author counted the views of the first 12 videos on the page themselves: total around 48.7 million views, with the highest video at around 20.4 million. The money figure is separated out. Some posts on X say the creator of Derek released a token on the pump.fun platform and made around $48,000 in transaction fees; another account said $60,000. The author does not count that as income: transaction fees from tokens are more like casino receipts than a salary, and they disappear when the hype dies down. Views sit right there on the page for anyone to check in 2 minutes.
Lesson 1: the face is not required, the silhouette is
Recognizable silhouette is the set of body features and clothing that help viewers remember a character even without seeing their face. According to GHOST (2026), most trend analyses about AI characters say the face is everything, but Derek almost always has his hair covering his eyes and is still the most recognizable character on the author's feed. For example, everything that makes up Derek sits below that hair:
- Black emo-style hair, long.
- Muscles so large they're absurd, become a joke.
- Same outfit in every video: gray shirt with blue graphic, studded belt, wallet chain, black cargo pants.
Viewers recognize him from the shoulders down, from behind, at the size of a postage stamp. The captions even add a paradox: "Should I be scary?", "I feel lonely..." from a sad teenager in a body that can flip a whole car. Viewers stay to see which side wins. The face might not be needed. The silhouette is mandatory.
Lesson 2: describe the character in one sentence
The one-sentence rule is a test that the character is done if the creator can describe it to a friend in one sentence. According to GHOST (2026), the author set this rule before writing the first prompt, and Derek passes the test in five words: emo hair, absurd muscles, gray shirt. That sentence is the product; everything else is decoration.
For his own character, the author chose four features large enough to survive at every camera angle: a hairstyle readable from across the street, eyes slightly too surprised, thin front teeth showing every time he smiles, and a suit in a color that acts as a logo. Add one more test borrowed from Derek's page: on a phone, each video square in a grid is only about the size of a thumb, and Derek is still recognizable at all 12 squares. So every character design must be recognizable when shrunk to the size of one grid square.
Process: Three Reference Sheets, One Character Retained
The creator builds the character using a character reference sheet: a single image on a smooth gray background, showing full body front view, side view, back view, plus two close-up face shots — so every video afterward uses the same "original version."
- First sheet — discard immediately. The character is strange, but deliberately so: hair like a hat placed on the head for a photo, a green velvet tuxedo looking like a wedding party staff member, a face tilted toward caricature. This kind of "strange" people scroll past, because it demands attention rather than earning it. The creator keeps three things: round surprised eyes, sparse teeth, and a square hairstyle.
- Second sheet — right face, wrong outfit. Built from scratch around the chosen features: black hair standing tall nearly equal to the face width, eyes showing white space around the pupils, sparse teeth with thin mustache. Face done. But the cream vest with black shirt and yellow tie "looks great" — and great is what nobody remembers. Shrink it down to a grid square, cream color blends into gray background, only the hair shows.
- Third sheet — edit, don't rewrite. This is the step that changes everything.
Why edit the image instead of writing a new prompt?
Editing an image means refining an existing reference sheet to keep the character intact and only change what you specify. A new prompt gives you a new person: even if you paste the old description exactly, the model still "rolls the dice" again and hands you the character's cousin.
The creator attaches the second sheet and requests keeping the face, eyes, teeth, mustache, skin, build, hair, background, lighting, composition exactly — only change the clothing, the same from every angle. The new outfit is turquoise silk vest woven with golden floral vine and bee patterns, pink silk shirt with large bow, gold chain and ring. The character is named Cornelius, because turquoise, pink, and gold is a combination nobody wears by accident. Shrink it down to a grid square, Cornelius is a turquoise block with a black hair column on top, recognizable before seeing any facial detail clearly, even from behind. The lesson packed in one line: design the face for close-up shots, design the outfit for thumbnail images.
Give the character a "shop" before giving him a joke
Derek lifts weights — that's his whole world: one gym, one balcony, one body built for exactly one joke. The creator wanted Cornelius to own something, because one business gives you hundreds of free videos worth of content. Cornelius becomes the owner of a luxury fur shop — not a sales clerk. That one word changes how he moves: a salesman tries to convince you, an owner just shows you what's his and waits for you to understand.
The first video script:
- Camera angle: phone placed on stone counter, tilted slightly upward — exactly like a real shop owner filming inventory for social media. From this angle, the tall column of hair under the ceiling fan and he looks twice the size of the room.
- Action: introduce each item straight into the lens with the pride of someone who hand-picked the merchandise.
- Setting: coat racks in cream color, honey color; two employees in tuxedos stand still in the background. Nobody reacts to the turquoise vest, because he's the boss — the world sees him as completely normal, and that's exactly what makes him strange.
- Final beat: he pushes a large white fur straight into the lens until it fills the frame, only his wide eyes peeking out from above — the only shot where the face nearly disappears, and it's the frame people screenshot.
There's one quieter reason too: someone introducing products in front of the lens is already an ad format. If a brand partnership comes later, the product placement already exists in the show — he picks up the product exactly the way he picks up the fur.
Where to Put the Reference Image: "reference", not "first frame"
The video was made using Seedance inside Higgsfield, with a full-length front-facing photo attached as a reference image, not as the first frame. The difference sounds technical but determines everything: use it as the opening frame, and the video opens with a studio photo on a gray wall and dies in half a second; use it as a reference, and the model takes the face and outfit into a completely new room.
Three sentences in the video prompt carry most of the result:
- "The phone never moves" — so the model doesn't make up a cameraman on its own.
- "The employee in the back stands completely still" — so the background doesn't grab attention.
- The brush stroke of hair into the lens at the end — gives the video an ending. Most AI videos don't end, they just stop.
With dance videos, I believe many videos in this wave take a viral dance already performed by a real person, then use motion transfer tools (for example Kling Motion Control) to make the character move identically. The body language is borrowed from a real person, so it looks real.
What does it cost to make 10 AI character videos?
The cost to make 10 complete AI character videos is around $155 according to the most detailed public cost breakdown that GHOST (2026) found: around $102 for the video part (320 seconds, around $0.32 per second), around $3.35 for 50 images, the rest for other AI calls, editing and music. Counting retakes, budget $200–300 for the first 10 videos.
The real cost is time. The same source estimates 30–50 hours for the first 10 videos while learning, and after spending an evening making a reference sheet, I found that number to be real. For example, if someone making content in Vietnam spends 2 hours a day, the first 10 videos take about 15 to 25 days, a calculation from OPCAIOS based on those estimates rather than my own data.
Is this AI character wave cooling down?
The AI character wave may be cooling down, according to the part that many articles about Derek overlook. According to GHOST (2026), Derek's early videos got around 2 million views each, while new videos struggle to break 100 thousand. The French character with a mustache that kicked off this whole wave is said to have lost about 50% engagement after viewers found out it was AI.
Novelty is the spark plug, not the engine. A weird appearance buys you the first million views and almost nothing after. What survives the spark plug is a habit viewers want to see again: the same person, same outfit, a small new situation each time. Cornelius's shop the following week sells hats, then perfume, then a very expensive pair of gloves. I also raised two unsolved risks: dense patterns like tiny gold bees that make the video model flicker easily when moving, and a formula easy to copy within a week if Cornelius succeeds. Build a program, not a moment.
Clearly mark "AI character" from the start
I've already decided: put "AI character" in the account introduction from the first post. Viewers click to learn how the character was made, and hiding it only makes comments angrier when they find out.
For content creators in Vietnam, this is also a matter of compliance: TikTok, YouTube and Meta all require labeling AI-generated content that looks realistic, and when the character starts getting ads, the ad content still has to follow the same advertising rules as any real KOL — being an AI character doesn't let you make exaggerated product claims.
If You Want to Try Making an AI Character, How Do You Start?
The roadmap for starting is to make 10 videos by hand around a single character. According to GHOST (2026), the author proposes five steps for themselves, from design to budget:
- One character, describable in a single sentence, recognizable at grid square size.
- One reference sheet locked in, every change after that is image editing, not writing new prompts.
- One repeating setting (store, counter, kitchen...) instead of a one-off joke.
- Ten videos made by hand, with a budget you can afford to lose without regret, around $200–300 by the estimate above.
- No second character yet, no automation yet, no tokens until viewers demand the next episode of the first character.
As the author writes, an AI agent can post 100 videos a day, but it can't fix a character that nobody wants to meet twice. After ten videos, you'll know whether a costume can become a character or not.
Read more: Building a "faceless" TikTok/Reels channel for OPC; Step-by-step creating an automated TikTok/Shorts channel using HeyGen, Suno and CapCut from A-Z.
Source: GHOST (@ghosstty_) on X, 16/10/2026 — rewritten and condensed by OPCAIOS; account figures and costs per original author.
Translated from Vietnamese with AI. Read the original.
Want to apply this in your business?
Book a free 30-minute call (Vietnam time, in English or Vietnamese).
