Hi tech & AI enthusiasts, this week's highlights:

  • Runway Gen-4.5 Tops Video Generation Leaderboard

  • Kling AI Launches First Unified Video Model

  • AI Shopping Traffic Surges 520% This Holiday Season

  • Plus: FLUX.2, Amazon $50B Government AI, Fujitsu Robot Collaboration

  • The Lab: How to Create Cinematic Grids With Nano Banana Pro

  • AI Image of the Week: Mustang cinematic grid

Let’s get into it.

AI MODELS

Runway launched Gen-4.5 on December 1, taking first place on the Artificial Analysis Text to Video benchmark with 1,247 Elo points and beating all competing video models. Watch the unveil here.

What you need to know

  • Accurate physics and motion: Objects move with realistic weight and momentum. Liquids flow properly, hair stays coherent during motion, and fine details don't fall apart between frames.

  • Better prompt following: The model creates what you ask for on the first try more consistently. It handles everything from realistic everyday scenes to complex multi-element compositions without needing multiple attempts.

  • Natural character expressions: Human subjects show nuanced emotions and lifelike facial details, though the model still struggles with effects happening before causes and objects disappearing unexpectedly.

Key Ideas

▲ Why it matters: Video generation models are getting good enough that creators can rely on them for actual production work rather than just prototyping. Runway's top ranking signals that AI video quality is reaching the point where it competes with traditional production methods for certain use cases.

AI MODELS

Kling AI released Video O1 on December 1, the first model that combines video generation, editing, and understanding in a single system rather than requiring separate tools for each task.

What you need to know

  • Generate and edit in one place: You can create videos from text, turn images into video, and edit existing footage using natural language prompts without switching between different tools.

  • Keeps subjects consistent: The model maintains character appearance, clothing, and visual details across multiple shots. When you upload reference images, it preserves those characteristics throughout the entire video instead of letting them drift.

  • Beats competitors on editing: In head-to-head tests, Video O1 performed 230% better than Runway Aleph on video editing tasks and 247% better than Google Veo on turning images into video.

Key Idea

▲ Why it matters: Video creators currently juggle multiple tools to handle generation, editing, and effects work. A unified system that does everything in one interface cuts production time substantially and eliminates the hassle of moving files between applications.

AI SHOPPING

Adobe's holiday shopping report shows that traffic to retail sites from AI services like ChatGPT and Perplexity is expected to surge 515 to 520% during the 2025 holiday season, with shoppers from AI now converting at higher rates than traditional channels.

What you need to know

  • AI shoppers buy more: By October, shoppers arriving from AI services were 16% more likely to complete purchases than people coming from search, email, or social media. They also spend 44% longer on sites and look at 12% more pages.

  • Mainstream adoption: 38% percent of Americans have already used AI for online shopping, with 52% planning to use it. People mostly use AI to research products, get recommendations, and find deals.

  • Record holiday spending: Online sales are projected to hit $253 billion from November through December, up 5.3% from last year. Mobile will account for 56% of that spending.

Key Idea

▲ Why it matters: AI has flipped from a research tool to an actual shopping channel in less than a year. Retailers that don't show up in AI results risk losing a growing chunk of holiday traffic to competitors who've optimised for AI discovery.

EVEN MORE NEWS

Other big ideas from the past week.

Created with Midjourney

FLUX.2 image model: Black Forest Labs released FLUX.2, their most capable image generation model yet. It renders at 4 megapixels and maintains consistent characters across multiple images.

Amazon government AI: Amazon announced up to $50 billion in investment to expand AI and supercomputing for federal agencies, adding 1.3 gigawatts of capacity starting in 2026.

Fujitsu robot collaboration: Fujitsu unveiled spatial world model technology that helps robots predict human behaviour and work safely alongside people in homes and offices.

The Lab

▲ AI EDUCATION

How to Create Cinematic Grids With Nano Banana Pro

Learn how to use Google's Gemini 3 Nano Banana Pro to generate professional-looking portrait headshots from a single photo of yourself.

What you'll learn:

  • How to create comprehensive 3x3 cinematic grids showing every essential angle

  • How to maintain perfect subject consistency across all nine panels

  • How to extract specific stills from your grid using simple prompts

  • Real-world applications for filmmakers, photographers, and content creators

🔬 Interested in learning about AI for business? Join The Lab waitlist.

AI IMAGE SHOWCASE

Create this type of image with Nano Banana Pro

A professional 3x3 cinematic storyboard grid containing 9 panels showing a vintage 1967 red Mustang convertible on an empty desert highway at golden hour. Row 1: Extreme long shot with car small on vast desert road, full long shot showing complete vehicle, 3/4 angle medium long shot. Row 2: Medium shot of front half of vehicle, medium close-up of front grille and headlights, close-up of front badge and details. Row 3: Extreme close-up of chrome details and reflections, low angle from ground level showing undercarriage, high angle bird's eye view from directly above. Same red Mustang, same desert highway location, same golden hour lighting, consistent cinematic color grading, photorealistic textures, proper depth of field with bokeh in close-ups, professional automotive photography.

Image Usage Suggestion: Automotive marketing, product photography reference sheets, or cinematography storyboarding

💌 Reply to this email with your AI image generation or suggestion of what you’d like to see and I'll feature it in a future newsletter.

SHARE YOUR THOUGHTS

Help shape the content you see here by giving feedback

You can add more feedback after choosing an option

Login or Subscribe to participate

Have specific feedback or want to get in touch? Reply to this email and I’ll get back to you.

Know someone who’d love this newsletter? Forward it to a friend and have them sign up here.

Thanks for reading, until next time.

Stay curious,
Matt Lok, Editor

Brought to you by Metalabs. Digital marketing consulting specialising in AI.