The 'How Different Our Lives Are' AI Trend, Explained

The “How Different Our Lives Are” trend puts the same person on screen twice, side by side, living two contrasting versions of the same moment — hair done and calm on one side, toast in mouth and sprinting for the door on the other. The whole joke lands because both halves sit in one frame at once. Flovaly has no dedicated preset for it, and — more importantly — neither of Flovaly's generation pipelines can reliably put the same character's face in two places in a single image or clip. This article explains why, and the honest two-step workaround that actually works.

What the trend actually is

Search results split into two genuinely different formats. The first, and the one most of the biggest TikTok posts use, is a real dual-clip video: the creator films the calm version and the chaotic version separately, then cuts them into a split screen with a standard video editor or CapCut's split-screen tool. No AI generation is involved at all — it's editing, not synthesis.

The second format is a single AI-generated still image. Tools built around this — Miraflow AI's Image Generator is the most visible one — take one uploaded photo and a prompt describing both halves at once, and render a single coherent frame with the subject appearing twice. Their own guidance is explicit about why that's hard: you have to “describe both halves of the split in the same prompt so the model generates them together in one coherent frame rather than as two disconnected images,” and matching lighting and camera angle across both sides is the most common failure point.

Why this doesn't work as a single Flovaly generation

Flovaly has two generation paths that could plausibly apply here, and neither is built for two instances of the same face in one output.

The Studio's Video tabtakes one prompt and, at most, one reference image as the start frame — there's no second clip, no end-frame field, and no split-screen layout option anywhere in the request the composer sends. It can only ever produce one continuous scene, not two contrasting ones stitched together.

The Studio's Image tabis closer in spirit to the Miraflow-style still-image version, but its mechanic is different in a way that matters here: it generates one new scene from your prompt, then swaps your character's face onto oneface in that scene. There's no setting that tells it “swap this face onto both people in the frame.” If the scene your prompt describes has the same person appearing twice, Flovaly's face-swap step has no way to paint your character's face onto the second instance — at best one side of the split will look like you, and the other will look like whatever face the image model happened to render for it.

The honest workaround: generate the two halves separately

Since one generation can't reliably deliver two convincing faces in one frame, the workaround mirrors the trend's own dominant video format — make two things, then combine them yourself:

1. Open Studio in the dashboard sidebar and switch to the Image tab.
2. Pick a saved character with a clear, front-facing reference photo — the same character for both generations, so both halves actually match.
3. Generate the first half: a prompt describing the “calm” version — composed, hair done, unhurried, in a specific setting (front door, kitchen, bathroom mirror). This costs 2 credits and uses Nano Banana Pro to build the scene, then face-swaps your character in.
4. Generate the second half separately: the same setting, camera angle, and lighting words in the prompt, but the “chaotic” version instead — half-dressed, toast in mouth, mid-sprint. Another 2 credits.
5. Combine the two images into one split-screen frame yourself, using any free photo editor or even a phone collage app — the same manual assembly step the trend's video creators already do with two filmed clips.

Getting the two halves to actually match

• Reuse the same setting words in both prompts. “Front door of a small apartment, morning light through a window on the left” repeated in both prompts gives the model the best shot at matching backgrounds — but since each generation is independent, expect some drift and pick two images where it reads close enough, rather than assuming a pixel-perfect match.
• Name the camera angle and framing explicitly. “Waist-up, straight-on” in both prompts reduces the chance one generation frames wider or from a different height than the other.
• Use a sharp, evenly lit reference photo.The face-swap step reads your character's saved reference image directly for both generations — a clear, forward-facing shot gives it more to match than an angled or low-resolution one.
• Budget for a few extra attempts.At 2 credits per half, regenerating one side to fix a mismatched angle or lighting is cheap — it's the combined frame that needs to look coherent, not any single generation.

What it costs

Two Image-tab generations at a flat 2 credits each — the standard Nano Banana Pro rate — comes to 4 credits total, roughly $0.78 on the $29 Starter plan's 150-credit allowance. See pricing for how credits map to every plan, or the full cost breakdown for how still images compare to every other generation type on Flovaly.

Where this fits

If you want the single-frame AI version with the least friction, a tool built specifically for that — one that generates both halves in one pass and handles the composite for you — will get there faster. Flovaly's honest angle is different: the same saved character you use for this workaround also carries over to every other format on the account — Mini Me, superhero portraits, a bob haircut preview, or a full talking-actor ad tomorrow — all from the same reference photo, on the same account as the rest of the Viral Centre line-up. It's a $1 trial to try the two-generation workaround yourself.

FAQ