My video walkthrough of the recent Midjourney 8.2 test.
Taking an existing still from a recent Seedance 2.5 piece I made —depicting a solo mountaineer in a harsh alpine environment — and running it through Midjourney's new v8.2 edit model to see how far I could iterate on a single frame without regenerating from scratch.
Tested wardrobe swaps, environment shifts, and style transfers.
The alpine mountaineer still is such a good stress test for v8.2, that much texture in the snow and gear usually falls apart under edit models. Did anything break or shift unexpectedly between iterations, or did it hold up cleanly the whole way through?
For this one, I went for a premium podcast-style look with natural delivery, subtle camera changes, kinetic text and supporting B-roll, instead of making it feel like the usual AI talking-head video.
The goal was simple: make AI-generated content feel like something you’d actually stop and watch.