EP 036 · 90 MIN SHIP IT HALF-BAKED, DRIP THE REST
Midjourney V7 is (Finally) Here! Draft Mode, Hot Takes, And A Mild Identity Crisis
The short version
Episode 36 is Fast Hours' same-day reaction to Midjourney V7's launch. Rory and Drew rank hundreds of personalization images live, test the new Draft Mode and voice prompting (including on mobile), and marvel at V7's texture detail, while noting Remix, Quality and Character Reference aren't live yet. Rory argues the incremental rollout beats Sora's overwhelming all-at-once launch, and both hosts share workflow tricks for turning a single ChatGPT image into scalable ad GIFs and on-brand icon libraries.
Key takeaways
- 01
Ranking personalization images changes results fast now: Drew ranked 700+ on launch day alone and planned to push toward 2,000 before judging V7.
- 02
Keep two personalization profiles if your taste is split: Rory runs a dark/gritty one and a soft/minimal one and blends between them.
- 03
Remix, the Quality parameter and the V6 upscalers were all still missing at V7's launch — check what's actually live before assuming a feature is broken.
- 04
Relax mode held coherence better than Turbo on longer, more complex prompts in early V7 testing, despite Turbo's speed advantage.
- 05
Build a scalable ad GIF by chaining small incremental edits in ChatGPT — zoom 0.5x, open the mouth 50% — rather than prompting the final frame directly.
- 06
Give ChatGPT a locked template (brand font name, hex code, 'in this exact style') to generate a whole on-brand icon library from one reference image.
- 07
Rory says Midjourney's incremental drip strategy beats Sora's all-at-once launch: shipping too many features together overwhelms users and buries the excitement in confusion.
Receipts · said on the record
“There's a few images I got where I'm seeing skin pores on the arm that are in the mid ground, not even in the foreground. I'm like, that's impressive.”
“If Sora just launched text to video and that's it, everyone could have got on board … Instead they launch this whole suite and everyone's like, what the hell do I do with this thing?”
“It's a really undervalued thought because that's kind of where we're at right now with all these tools and all these options. Your one shining moment literally lasts 15 minutes.”
“Cause everybody uses the same icons. You know, so you're thinking about like, how do you be different? How do you add personality to your brand?”
“I saw it and I'm like, are any of these companies asking if consumers want to listen to an AI avatar? Does anyone want that?”
Questions this episode answers
What was missing from Midjourney V7 at launch?
Remix, the Quality parameter and updated upscalers weren't live yet — V7 still used the old V6 upscalers — and character/object reference (what would become Omni-Reference) hadn't shipped, with Midjourney's roadmap targeting new features roughly every one to two weeks over the following 60 days.
Do I need to re-rank my personalization profile for Midjourney V7?
Yes. Rory and Drew both found their old V6 personalization codes carried over but produced 'jacked up' results on V7, so they re-ranked from scratch — Drew rated over 700 images on launch day alone and planned to push toward 2,000.
Is Relax mode or Turbo mode better for Midjourney V7 image quality?
In Rory's early testing, Relax mode held together longer, more detailed prompts more coherently than Turbo, which felt less 'sticky' to the prompt despite being faster, similar to how some video models trade quality for speed in their fast modes.
How can I turn one ChatGPT-generated image into a set of ad creative?
Rory chains small incremental edits in the same ChatGPT thread, for example zooming in 0.5x and opening a mouth 50% per step, to generate a sequence he stitches into a looping GIF. He uses a similar locked-template approach (brand font name, hex code, 'in this exact style') to generate a whole on-brand icon set from one reference image.
What do Rory and Drew think of AI avatar tools?
Both are openly negative on them. Rory says he doesn't understand who actually wants to listen to an AI avatar, and Drew calls the idea one of the least interesting things he's seen, though they concede the underlying animation and lip-sync tech could be priming audiences for other applications.
Show notes
Fast Hours Podcast Ep. 36—Midjourney V7 is (Finally) Here! Draft Mode, Hot Takes, And A Mild Identity Crisis (with Drew Brucker & Rory Flynn)
In this episode of Midjourney Fast Hours, Drew Brucker and Rory Flynn react to the brand new Midjourney V7 update—and things get weird (in the best way). Draft mode? Feels illegal. Personalization? Feels personal. Voice mode? Sounds like sorcery. They throw around hot takes, half-baked theories, and the occasional “wait, what just happened?” as they pick apart V7’s weird new powers and what it means for AI creativity.
Somewhere between icon design tangents and ChatGPT hacks, they stumble into deeper questions about taste, tech, and whether artists are being replaced—or just upgraded. It’s messy, curious, and maybe smarter than it sounds.
▶️ What you'll learn in this video: • First impressions of Midjourney V7 • How Draft Mode compares to Turbo Mode • Voice Mode and the future of AI on mobile • The impact of V7 on image quality and creative workflows • ChatGPT x Midjourney integrations for better prompts • Use cases for icon design, marketing visuals, and AI avatars
👇 Drop your thoughts in the comments: What feature in V7 has you the most excited (or confused)?
