CREATIVE CIVIL WAR + MINIMAX IS HILARIOUSLY AWESOME.

Fam,

The AI landscape is getting more polarizing.

It was to be expected with the speed of development.

And it’s not slowing down.

Here’s what we got for today:

1. RUNWAY: RELEASES KEYFRAMES.

2. MINIMAX: HILARIOUSLY AWESOME AI VID GEN.

3. CUSTOM AI WORKFLOWS: NOW AVAILABLE.

4. PODCAST: MJ AFTER DARK.

5. KREA: THE MULTI-MODAL SOLUTION.

6. BEYOND MJ: NEWS ROUNDUP.

7. MJ OFFICE HOURS.

LET’S BOOGIE…

RUNWAY: RELEASES KEYFRAMES

Last week Runway joined the Keyframes party.

It was a much-needed edition anddd…of course we tested it.

(What I didn’t expect…the creative civil war in the comments section)

IF YOU MISSED IT → FULL POST

WHAT IT IS:

• You input the first frame the video

• You input the last frame of the video

• RW generates everything in between

PROCESS:

1. Create images in Midjourney.

2. Open Runway Gen-3

3. Upload a first frame image

4. Upload a last frame image

5. Add a text prompt

6. Generate + Iterate.

RUNWAY PROMPTS:

SCORPION → panning left shot, scorpion monster walking eerily towards as the camera focuses for a closeup

TESLA SKETCH→ hand drawn sketch comes to life quick, tracking shot of a tesla cyber truck driving left on a desert road environment

CYBER TRUCK ORBITY → camera rotates left around a tesla cyber truck in a dark studio environment

HAMBURGER → static shot, hand placing a bun atop of a juicy hamburger, natural dynamic motion, 35mm style

Here’s an example below ↓

MINIMAX: HILARIOUSLY AWESOME AI VIDEO GENS

If you haven’t heard of Minimax (by Hailuo)…definitely check it out.

It just stepped out of BETA, but there’s a free trial.

It’s has more complex motion than most video generators, but it’s still raw.

CHECK IT → FULL POST

WHERE IT WINS:

• The motion is more natural and fluid.

• More complex motion is possible.

• It has the fastest motion of all the vid generators.

• It’s improving quickly.

WHERE IT LOSES:

• There’s a lot of glitching/morphing

• Typically long generation times

• Text prompts are ok, but not super consistent

CUSTOM AI WORKFLOWS:

We’ve been cooking in the lab at SUPERSIDE.

Seems like each week we’re working on something new…and most of the time, these custom workflows end up becoming productized.

Like this “Breif-to-Email Marketing Campaign” workflow for Vimeo ↓

WHY?:

• So we get from brief-to-first-draft as rapidly as possible.

• With key visuals/copy/ux/ui trained and tailored to brand guidelines/standards

• Fully automated with tools like OpenAi/Claude/Flux/Figma

• So we cut the back-and-forth time on “does this fit the brand?”

• And we focus on strategy, visual firepower, and iteration.

Have a design problem that needs solving? (We like those)

It’s worth a convo (at the very minimum).

KREA: THE ALL-IN-ONE PLATFORM

I’ve always enjoyed KREA’s platform…but they’re making a real interesting play.

Instead of owning the real-time gen market…they’ve integrated multiple API’s.

So if you want to try all the video tools (Runway/Kling/Minimax/Luma)…for one price…

CHECK OUT → KREA

WHY IT’S IMPORTANT?

• You can generate images and video on the platform

• You can also upscale and edit generated or existing images 

• People DON’T WANT to pay for 30 platforms

• Consolidating them into a simple ux/ui is super helpful

• Their pricing is reasonable enough to access all these tools.

It’s also cool you can upload different LORAs and try ones from the community. I think this was a really smart move and great way to bring more people into the environment.

PODCAST: MIDJOURNEY AFTER DARK

Drew and I are back with the 2nd edition of MJ After Dark.

A few drinks…a couple weird tangents…an overall good time.

HIGHLIGHTS:

• [00:00] Intro | [05:45] AI Video Wars | [12:08] Platform Comparisons | [21:24] Style Code Exploration | [26:44] Aspect Ratio Effects | [32:37] Creative Process | [35:28] Community Curation | [39:28] Text Improvements | [42:00] Old Prompt Revival | [44:54] Future Possibilities

If you like the episode…drop us a “LIKE” or “SUBSCRIBE” (it really helps)

Respond to this email if you have any feedback (it’s all helpful).

MIDJOURNEY OFFICE HOURS (10/9):

New Editor

• Import and edit external images

• Depth-sensing for 'filter' features and inpainting

• Initially for Pro/Mega yearly subscribers only

• Web-only, separate from general 'Create' workflow

Personalization Features

• Cold-start new personalizations

• Fast and easy personalization creation

• Multiple personalizations from scratch

• Possible 'style personalization' feature

Workflow Changes

• UI updates with default personalization

• Option for 'balanced' style

• Possible 'Batch8': 8-image grids, wider range of styles

Version 7

• Bug fixes, currently in training

• Better multi-lingual capabilities

• Improved prompt understanding and image coherence

• Better base model for new and existing features

Video/3D

• Good but slow/expensive model developed

• Aiming for clean, fast, and affordable release model

• Potential for specialized video models later

• First major dataset collected, training to begin

• Release timing (before/after video) still debated

Web UI Enhancements

• Adding Prefer Option abilities

• Better organization features (subfolders/tags)

App Development

• Web and Mobile apps in development

Have fun.

If you need help…DM me on Linkedin.

Until next time…

Rory