Episode
Midjourney Edit Models + Nano Banana Pro: Eating Lunches in the Upside Down
- Podcast
- Fast Hours
- Published
- Dec 3, 2025
- Duration seconds
- 4097
- Processing state
not_requested
Actions
POST https://stenobird.com/v1/public/podcasts/fast-hours-6954465/episodes/midjourney-edit-models-nano-banana-pro-eating-lunches-in-the-upside-down/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/fast-hours-6954465/midjourney-edit-models-nano-banana-pro-eating-lunches-in-the-upside-down.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Rory and Drew celebrate crawling their way to 30k subs, then immediately prove they are barely qualified to handle it by turning a Stranger Things binge into a full-blown lecture on composition, lighting, and how to reverse-engineer blockbuster shots into Midjourney and Nano Banana Pro prompts. They talk like film school dropouts who discovered prompts instead of lenses. From there, they unpack fresh Midjourney office hours: the upcoming UI/UX overhaul with continuous scrolling, better color control, a reworked style system, and the big one: parallel edit models that finally keep you inside Midjourney instead of forcing you into five other tools. They break down what “better text handling” could realistically mean for real-world client work, what to expect from Midjourney V8 training in January, and why business use cases will decide who actually wins this model war. Then it’s a long, dangerous slide into Nano Banana Pro obsession. They show how they are using it for real campaigns: ingredient flat-lay diagrams with perfect labels, knolling that actually respects object counts, thumbnail iterations in minutes, hyper-real food tweaks (“make the cheese more brown and bubbly”) and product work where text on bottles and labels actually holds up. Think: turning moodboards into branded cars, movie-poster typography onto existing art, and multi-shot car sequences that are clean enough to use as video keyframes. In the back half, they zoom out into systems: building custom Nano tools in Google AI Studio, using JSON prompts, if-then logic, and style libraries to create reusable pipelines for teams that are not prompt nerds. They rant about broken N8N workflows, fake Instagram “AI automation” grifts, and share where affiliate tools actually see conversions today across YouTube,…