WTF RUNWAY? + NEW PRODUCT SHOT AI TOOL.

Fam,

No intro needed…(i’m too tired)

Here’s what we got for today:

1. RUNWAY: ANIMATING YOURSELF

2. MIDJOURNEY: PERSONALIZE UPDATE

3. RUNWAY: UPDATES + GEN48 WINNERS

4. AI AD CREATIVE SYSTEM: AGENTS + KNG. GRAPHS.

5. PODCAST: YOUR BURNING QUESTIONS

6. BEYOND MJ: QREATES IS NOW AVAILABLE

LET’S BOOGIE…

RUNWAY: ANIMATING YOURSELF

Well this was wild…

I recently received the video of my presentation in Budapest. It’s awesome quality…but majority of it is shot from long-distance. (And I needed some closeups for B-roll)

So I used Runway to animate still images (of myself).

IF YOU MISSED IT → FULL POST

PROCESS:

1. Collect Hi-Res Images

2. Upload them into Runway Gen-3

3. Select “First Frame”

4. Add a text prompt

5. Generate & Iterate

RUNWAY PROMPT STRUCTURE:[SHOT TYPE], keynote speaker [action], natural dynamic motion, 35mm style

RUNWAY PROMPT EXAMPLE 1: side profile, keynote speaker talking to an audience on stage, walking to the right, natural dynamic motion, 35mm style

RUNWAY PROMPT EXAMPLE 2: Panning left shot, keynote speaker talking to an audience on stage, natural dynamic motion, 35mm style

The animation is cool, but the process itself is useful.

Now, that we have the clips…I added as B-Roll to a full reel.

Here’s an example below ↓

This is fun…try it out for yourself and LMK how it works.

More to come.

RUNWAY: NEW FEATURE UPDATES

So Runway also just dropped some new updates.

Most of these are for GEN-3 ALPHA TURBO (meaning everything just got faster)

1. GEN-3 ALPHA TURBO: Video-2-Video now available

2. GEN-3 ALPHA TURBO: Vertical Video (9:16) now available

3. EXTEND: IMG-2-VIDEO gen extension now available

4. EDIT: Trimming your clips is now available

RUNWAY GEN48

Runway also announced the winners of their GEN48 Film Festival.

Some of the winning submissions were incredible. Generating these shorts takes a ton of time and a lot of talent…and the winner submissions were worthy.

It’s 100% worth checking out the winners → GEN48 Winners

I pulled some of MY FAVORITES here as well.

MIDJOURNEY: PERSONALIZE UPDATE

The Personalize feature continues to evolve.

And Midjourney just made it more of a priority to tune your style.

(And they added a “Personalize” tab)

PRETTY SIMPLE:

The more images you rank…the more your style codes will be personalized.

This is also a great way to get “FREE Fast Hours”.

As always…you don’t have to rank every image…rank the ones you really like as it will affect your Personalize code.

AI DESIGN SYSTEM: SOLVING BIG ISSUES

This stuff we’ve been producing with Superside still blows my mind.

If you need a visual of how this beast works. (Check it out below)

KNOWLEDGE GRAPHS + AGENTS:

We hit a lot of roadblocks before this was the structure.

Instead of training individual agents for every client (500+)…we trained the agents to read knowledge graphs and produce based on them.

WHAT IS A KNOWLEDGE GRAPH? → Essentially it’s a super detailed brand guideline…and an operating plan for the agents…so they know where to look…and how to produce.

Using the Knowledge Graphs and the Agents together…allows us to produce for multiple brands simultaneously.

» If you have design/creative needs…whether that’s Ads, 3D, Web Dev, Rapid Branding & Concepting, Social Media, Video Production, etc «

PODCAST: AMA - YOUR BURNING QUESTIONS:

Really killer pod this week. (loved doing this one)

You asked your questions…and we dialed up some answers.

IF YOU MISSED IT → WATCH HERE

EPISODE HIGHLIGHTS:

• [00:00] Intro

• [02:54] Photorealism Techniques

• [14:15] Character Consistency + --cref

• [21:15] Mood Boards

• [32:55] Brand Styling

• [43:41] Body Proportions

• [50:25] Upscaling Methods

• [53:55] Creative Upscaling

• [58:58] Industry Developments

If you like the episode…drop us a “LIKE” or “SUBSCRIBE”

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

BEYOND MJ: QREATES IS NOW AVAILABLE

DTC/E-Comm folks…this ones for you.

QREATES is aiming to solve the problem of Ai product photos.

I took it for a test drive and it works pretty well.

CHECK IT OUT.

PROCESS:

1. Upload a product photo (png recommended)

2. Choose or upload a reference image

3. Add a text prompt

4. Generate product photos

THOUGHTS:

• It keeps branding very consistent

• The reference image influences the output a lot

• The ref image feature is actually really nice.

• You can generate backgrounds in Midjourney + add to creates

• Use Higher-Res product shots for better results.

Going to test this one way more…stay tuned.

MIDJOURNEY OFFICE HOURS (9/25):

Video Models Development:

• Weekly improvements in training video models

• Considering V6 video model pre-alpha release

• Debating native resolution vs. upscaling process

• Training multiple versions to assess approach

3D Model Progress:

• Daily collection of 3D data underway

• Larger tests expected after a week of collection

• May be video model extension or independent

• Release timing: before or after video model

Release Strategy and Lessons:

• V7 expected to take longest due to changes

• Early V8 discussions for smaller, manageable updates

Storytelling Team Updates:

• First demos and interactive releases coming soon

• Sessions planned to showcase tools and get feedback

• Tools are experimental, not long-term products yet

• Intent to evolve based on user feedback

Have fun.

If you need help…DM me on Linkedin.

Until next time…

Rory