- Uncovering AI
- Posts
- 📸 Video models just changed forever (Kling’s 5-Day Sprint)
📸 Video models just changed forever (Kling’s 5-Day Sprint)
Is Google’s new $250/mo model worth it? Plus: How to edit anything inside a video.
My fellow AI explorers
I’ve been glued to my screens tracking a five-day marathon of announcements from Kling, a surprise tease from Runway, and Amazon taking over Las Vegas with re:Invent. It feels like the industry decided to launch everything at once before the holidays kick in.
From "edit anything" video models to Amazon’s army of agents, the pace isn't slowing down. This edition is packed with the fun, the interesting, and the weird.
In today’s edition:
🎥 The Video Wars: Kling’s 5-day sprint vs. Runway’s Gen 4.5
🏠Amazon re:Invent: AI Factories and Frontier Agents
🧠Google’s DeepThink: Expensive, smart, and building agents
🌤️ 30-Second Play: The "Sim City" Weather Trend
Must See AI Tools
đź’° Payman: AI That Pays Humans. Over 10,000+ signed up for the beta
💫 SubMagic: An AI tool that edits short-form content for you! (Get 10% off using code “uncoverai” at checkout)
🎤 11Labs: #1 AI voice generator (Click Here to get 10,000 free credits upon signing up!)
🤖 ManyChat: Automate your responses & conversations on IG, FB and more! (Click Here to get first month for free)
🎙️ Syllaby: The only social media marketing tool you’ll ever need - powered by AI! (Get 25% off the first month or any annual plan with code “UNCOVER” at checkout)
AI Video
🎥 The Video Model Explosion: Kling vs. Runway vs. Apple
It wasn't just one announcement this week; it was a barrage. The Chinese model maker Kling AI went on a five-day spree, dropping update after update. But they weren't the only ones—Runway and even Apple decided to join the party.
Here’s the breakdown:
Kling 01 & Image 01: Described as the "Nano Banana" of video, this allows you to edit almost anything within a video or image using text prompts. (Think changing a book to a glowing tablet while keeping the hand movement perfect).
Kling Video 2.6: Finally, native audio. No more generating video and pasting sound later. It attempts to sync sound effects (like footsteps on leaves) directly with the visuals.
Runway Gen 4.5: Runway teased their new model, and honestly? It looks insane. The leaderboards suggest it might be blowing everything else out of the water with photorealistic consistency.
Apple StarFlow V: A surprise entry! Apple demoed a "normalizing flow-based" generator. It’s a different tech stack than the usual diffusion models, promising faster end-to-end training.
The Verdict? We are inching closer to Hollywood-level control. While Kling’s "edit anything" feature is hit-or-miss (it struggled with my melting cheese pizza car), the ability to swap objects in video while retaining motion is a massive leap.
🔮 Prediction: The next battleground isn't just generation quality—it's control. Expect 2025 to be the year where "text-to-video" becomes "text-to-edit," allowing precise manipulation of scenes rather than just rolling the dice on a new generation.
Founders…
🤖 Need an AI Agency to Help Your Business Implement AI Solutions?
Our preferred partner Align AI provides you with an expert AI and Automation implementation team to add 10-40+ hours of increased productivity per employee and achieve your goals faster.
Amazon
While the creative world focused on video, Amazon took over Vegas for re:Invent to focus on the enterprise. They aren't just building chatbots; they are building the infrastructure for autonomous workers.
Key announcements:
Frontier Agents: AWS unveiled a new class of agents, including a Developer Agent (writes/fixes code), a Security Agent (finds vulnerabilities), and a DevOps Agent (operational triage).
Trainium 3: A new AI chip designed specifically for training on AWS. It’s Amazon’s answer to Nvidia, optimizing cost and efficiency for their own ecosystem.
AI Factory: Amazon is bringing the cloud to you. This service allows companies to run dedicated AI infrastructure (inference and training) on-premise, using Amazon’s stack.
Why this matters: Amazon is positioning itself as the "operating system" for enterprise AI. By offering the chips, the data centers (AI Factory), and the employees (Frontier Agents), they are locking in big business.
đź’¬ Use case: Imagine a DevOps workflow where you don't get woken up at 3 AM. The DevOps Agent detects the crash, triages it, and suggests the fix before you even open your laptop.
AI SaaS Founders
🚨Want Millions of Impressions For Your AI SaaS, Done For You?
At uncovernews.co, we specialize in getting AI SaaS products the attention they deserve through strategic influencer marketing campaigns designed to drive millons of impressions at the fraction of the cost!
Get Your AI Startup’s News or Product In Front of Millions Quickly
🧠Google’s Agent Builder & DeepThink
Google quietly rolled out Gemini 3 DeepThink, their smartest reasoning model yet, but there's a catch—it’s currently gated behind the "Ultra" plan (approx. $250/month).
What’s new:
Gemini 3 DeepThink: Crushes benchmarks on coding and reasoning. It’s capable of fleshing out complex game design documents in a way previous models couldn't.
Workspace Studio: Think of this as "Zapier powered by Gemini." It’s a no-code agent builder inside Google Workspace.
Logic Flows: You can build triggers like: "When I get an email, check if it has a question. If yes, label it 'To Respond' and ping me in Chat."
The Reality Check: Access is tricky. Even with the expensive plan, the model can get overloaded ("I need a moment to sort through all those deep thoughts"). But when it works, it’s a powerhouse for logic-heavy tasks.
🌍 Takeaway: The barrier to entry for building complex AI agents is vanishing. You no longer need to know Python; you just need to know how to describe your workflow in plain English inside Google Workspace.
30-Second AI Play
🌤️ Trend Alert: The "Sim City" Weather Report
There is a new aesthetic trend taking over social media: cute, isometric, Sim City-style weather reports generated by AI. It’s weirdly satisfying and a fun way to use image generation.
Here’s how to do it:
Open your favorite image generator (Works best with Flux or Nano Banana Pro).
Use this prompt structure: "A cute isometric 3D render of [CITY NAME] with live weather conditions. Miniature tilt-shift style, detailed buildings, [WEATHER TYPE: e.g., rainy, sunny, snowy], sim city aesthetic."
Hit generate.
Why it’s cool: It turns boring weather data into a piece of digital art. People are using it to share their local vibes on social media—from a rainy London to a dusty Austin, Texas.
đź’ˇ Pro tip: Add "night mode" or "neon lights" to the prompt for a cyberpunk city vibe.
Other Relevant AI News!
🇫🇷 Mistral releases the Mistral 3 family, a powerful set of open-weights models under the Apache 2.0 license, making them free for anyone to use and fine-tune.
🚀 DeepSeek shocks the industry again with Version 3.2 and 3.2 Special, offering reasoning capabilities that rival the big US labs at a fraction of the training cost.
🧄 Rumors are swirling about OpenAI's "Garlic," a codename for their next big model (possibly GPT-4.5 or 5) rumored to drop early next year—or maybe even next week.
📢 OpenAI is testing ads in ChatGPT, with users spotting "Search Ads" for Target appearing directly in the interface—thankfully not inside the AI response text yet.
🎮 Epic Games CEO Tim Sweeney calls on Steam to drop the "Made with AI" warning labels, arguing that soon all games will use AI in some form. Read More.
đź“ş Amazon's Alexa Plus will soon let you describe a movie scene to jump straight to it on Prime Video, saving you from fast-forwarding hell. Check it out.
Golden Nuggets
🎥 Video AI is shifting from "generation" to "editing"—control is the new king.
🤖 Agents are the new Apps. Google and Amazon are racing to build the platforms where these agents live.
🔓 Open Source isn't dead. Mistral and DeepSeek are proving you don't need billions of dollars to build state-of-the-art models.
Would love to hear your thoughts on the news! Send me your thoughts by replying to this email (yes, I read them all :)
Until our next AI rendezvous,
Anthony | Founder of Uncover AI