Friday 🧪Lab Notes: The Rise of the Image Jockey

Hey MightyBrawnies and 💪MightyPros,

It’s already February. Quick reminder: our January Megadrop is live and ready. If you haven’t grabbed it yet, start here:

Megadrop post: https://brawny.ai/january-megadrop-1590-items-packed-and-ready/

Now for a fun Friday-night update.

Over the last few days I’ve been testing two open-source image editing models that genuinely change how fast I can work:

• Qwen 2511 (Edit) — from the Alibaba team
• FLUX.2 Klein 9B (Edit) — from Black Forest Labs (the same lineage that helped kick off the early Stable Diffusion era)

What’s new isn’t “making images from scratch.” It’s the ability to take an existing image and remix it on command—swap details, tweak styling, adjust outfits, refine poses, keep the vibe… without rebuilding everything from zero.

Three years ago I coined a term for this: the Image Jockey.

Same idea as a DJ, but for visuals: remix, variation, continuity, speed—turning one “seed” image into a full set that supports storytelling.

If you want the original definition + the early write-up from 2023, it’s here:

Image Jockey article: https://brawny.ai/image-jockey-ij-a-new-age-of-digital-artistry/

Why this matters for BrawnyAi

A lot of “one-click” mainstream tools can do these edits easily—but they’re also built around strict content rules. BrawnyAi relies on open-source so the creative toolkit stays sovereign long-term (and doesn’t disappear overnight because a third party changes the rules).

Open source still takes more setup and tuning, but the upside is huge:

• faster character consistency
• faster story-building
• reviving older projects that used to be too painful to produce
• less dependency on any single platform

What you’ll see in today’s gallery

In this post I’m showing examples made with Qwen 2511 and FLUX.2 Klein 9B Edit—including a small “mini-sequence” where the same couple travels through Rio in a consistent style.

Here’s a tighter, smoother version you can drop in, with clear “slots” for where you’ll place each image/gallery. I kept it simple and human.

Walkthrough (quick + visual)

Clean the sourceFirst, I remove anything I don’t want to carry over. In this case, I strip the original background and start from a clean white base.

Direct the editThen I tell the model what to change using a prompt (and sometimes an extra reference image if I want tighter control).

Build consistency Once the character holds steady, you can “chain” variations and even pair two characters to create a mini story sequence.

Results: Qwen 2511 EditHere’s the same setup rendered with Qwen 2511 Edit.

Results: FLUX.2 Klein 9B EditAnd here’s the same idea rendered with FLUX.2 Klein 9B Edit.

Which one wins for you: Qwen or FLUX?

Reply and tell me what you’re judging on: face, anatomy, pose stability, lighting, or overall vibe.

If you’re logged in as a MightyBrawny / 💪MightyPro, you’ll also see extra sets that push consistency further, so you can judge the results for yourself (gallery at the end of this post).

If you’re the kind of creator who wants to explore this universe, the workflow I’m using is built in ComfyUI. It’s a powerful interface for running open models locally. More info here:

ComfyUI: https://www.comfy.org

That’s it for tonight. Have a great weekend, and I’ll be back soon with more experiments—and more hunks.

With love,

B

Gallery for 💪MightyPros & MightyBrawnies!

Access additional images and videos below!