Sunday Rundown #121: Kling Is King & Paper-Paper-Paper
Sunday Bonus #81: Tool that generates a marketing kit from product image + description.
Happy Sunday, friends!
Welcome back to the weekly look at generative AI that covers the following:
Sunday Rundown (free): this week’s AI news + a fun AI fail.
Sunday Bonus (paid): an exclusive segment for my paid subscribers.
In case you missed it, here’s this week’s Thursday deep dive:
Note: If you’re consistently missing out on my emails, remember to check your “Promotions” tab and mark whytryai@substack.com as a “Safe Sender.”
Let’s get to it.
🗞️ AI news
Here are this week’s AI developments.
👩💻 AI releases
New stuff you can try right now:
Amazon news:
Frontier Agents is a suite of three autonomous agents like Kiro that can code and triage with minimal supervision.
Nova 2 is a multimodal model family, along with Nova Forge to train custom versions and Nova Act for building and managing AI agents.
Anthropic introduced Claude for Nonprofits, granting them a 75% discount on AI usage, connectors to popular nonprofit tools, and a free “AI Fluency” course.
ByteDance launched Seedream 4.5, an image model with better subject consistency and small text rendering.
DeepSeek open-sourced V3.2 and V3.2-Speciale: two reasoning models built for agentic tool use.
Google news:
Gemini 3 and Nano Banana Pro in AI Mode expanded to nearly 120 countries, unlocking better reasoning and multimodal answers.
Gemini 3 Deep Think is a stronger reasoning mode for Google AI Ultra users that handles tough math, science, and logic problems.
Workspace Studio gives business users a no-code way to build Gemini‑powered agents that automate work across Google products.
Kinetix launched Kamo‑1 in open beta, a video model that lets you control both full-body motion and camera movement for cinematic results. (Try it here.)
Kling AI news (omni launch week):
Avatar 2.0, a talking character model that supports up to 5-minute videos with more accurate gestures and facial expressions.
Image O1 image model can seamlessly blend up to 10 reference images, make edits based on simple commands, and mimic styles.
Video 2.6 is the company’s first native video model that generates sound effects and dialogue to go with the video.
Video O1 is a multimodal video model and platform that lets you generate and make precise edits to existing clips via simple commands.
KREA AI launched Node App Builder that turns your visual workflows into simple apps for others to use.
Mistral AI open-sourced Mistral 3, a family of multimodal models including Mistral Large 3 that boasts top-tier multilingual and reasoning performance.
OpenAGI launched Lux, a foundation computer‑use model that performs well on real-world, web-based tasks according to the Mind2Web benchmark.
Runway released Gen‑4.5, a new text‑to‑video model that promises SOTA motion quality, prompt adherence, and visual fidelity.
🔬 AI research
Cool stuff you might get to try one day:
Apple is working on STARFlow‑V, a normalizing flow‑based video model that matches diffusion models in quality.
OpenAI is reportedly working on a new model codenamed “Garlic” to regain its lead in reasoning and other benchmarks. (Possibly to launch as GPT-5.2?)
📖 AI resources
Helpful AI tools and stuff that teaches you about AI:
“How AI is transforming work at Anthropic” [ARTICLE]—real-world insights based on a survey of 132 Anthropic engineers.
“What 1,250 professionals told us about working with AI” [TOOL & STUDY]—findings from Anthropic’s new “Interviewer” project about people’s use of AI.
🔀 AI random
Other notable AI stories of the week:
OpenAI was rumored to be gearing up to introduce ads on ChatGPT but reportedly scrapped those plans due to “code red” over Google’s lead.
🤦♂️ AI fail of the week
I love a game where everyone’s always a winner.
Send me your AI fail for a chance to be featured in an upcoming Sunday Rundown.
💰 Sunday Bonus #81: Turn a product image + description into a full marketing kit
This week, I wanted to push my go-to Google Opal app maker by creating someting that calls on most of its features.
The result is Product Marketer, a tool that takes your product image and description and generates a marketing kit, complete with:
Product copy
List of key selling points / benefits (bullet list)
Taglines and slogan ideas
Social media post ideas
Marketing campaign ideas
Visuals for a product / lifestyle shot
Short video teaser for the product
It should go without saying that no AI tool will ever give you ready-to-go marketing collateral from scratch, but Product Marketer is a solid way to get some inspiration and overcome the “blank page” problem.