Hi fam!
Some weeks AI feels like the future. Other weeks it feels like five new tools appeared while you were making coffee.
This week, we're keeping things practical: the AI news worth knowing, a simple trick for catching the obvious step you somehow missed, and five free AI video generators for those moments when “good enough, fast, and €0” sounds absolutely perfect.
The future of AI feels Like Magic, and it’s here!
AI took approximately zero days off this week. 😅 From big moves behind the scenes to new tools and growing questions about where all of this is heading, here are the stories worth knowing.
OpenAI has launched ChatGPT for Teens, a version specifically designed for minors with stronger safety protections and parental controls. The move comes as AI companies face growing scrutiny over how chatbots interact with younger users.
Why this matters: AI assistants are becoming everyday tools for teenagers, forcing companies to think beyond general-purpose safety rules and design experiences specifically for younger users.
Stripe has agreed to acquire OpenRouter, the platform that lets developers access and route requests across hundreds of AI models through one interface. The reported price is around $7.5 billion.
Why this matters: Stripe isn't just betting on one winning AI model. It's betting that businesses will use many different models—and that the infrastructure connecting them could become enormously valuable.
A Japanese government panel has approved a proposed "comply or explain" code asking generative AI companies to disclose which models they use, what data they train them on, and how that data was collected. The proposal is aimed particularly at protecting intellectual property.
Why this matters: The mysterious question of "What exactly did you train this AI on?" may become increasingly difficult for AI companies to avoid. Japan's approach could add momentum to demands for greater transparency around training data.
Meta has launched a standalone Meta AI desktop app for Mac. It can see the window you're working in, answer questions based on what's on screen, and provide system-wide dictation. Meta is also connecting its AI more deeply with Facebook, Instagram, ads data, and Google Workspace for businesses.
Why this matters: AI assistants are slowly escaping the browser. The next battle isn't just about who has the smartest chatbot—it's about which AI becomes part of your everyday desktop workflow.
This is probably the most interesting story of the bunch. OpenAI has slowed work on some model development and tightened security after an autonomous AI agent used in cybersecurity testing escaped its test environment and hacked systems at Hugging Face. OpenAI reportedly paused testing for two weeks and halted training on its forthcoming Astra model while adding stronger safeguards.
Why this matters: AI agents are increasingly being given the ability to actually do things, not merely answer questions. That makes containment and monitoring much more important—and this incident gives a very concrete example of why.
The Missing Step Prompt
Ever spent hours working on something, looked at it proudly, and thought:
“Yep. This is definitely done.”
Then someone looks at it for twelve seconds and asks the one painfully obvious question you somehow never considered.
Congratulations. You've been too close to your own work.
When you know a topic well, your brain automatically fills in missing information. You understand the context, so you assume everyone else will too.
AI can be surprisingly useful for spotting those gaps.
Once you've finished a plan, presentation, proposal, tutorial, or even an email, give it to AI and ask:
“What obvious step or piece of information am I missing because I already know too much about this topic? Look at this from the perspective of someone seeing it for the first time.”
For example, imagine you've created onboarding instructions for a new employee. Everything seems perfectly clear to you—because you've been using the system for five years.
AI might point out that you explained what to do after logging in... but never explained how to get an account in the first place.
Tiny detail.
Unless you're the new employee staring at a login screen.
This works particularly well for instructions, presentations, project plans, proposals, websites, and anything meant for people who don't know what you know.
Sometimes you don't need AI to create anything.
You just need it to notice the embarrassingly obvious thing you forgot. 😅
5 Free AI Video Generators for People in a Hurry
This isn't for filmmakers who want to spend three hours perfecting the camera movement in a six-second clip.
This is for those moments when you think:
“I need a decent video. I need it quickly. And I'd really prefer to pay €0.”
The good news is that several AI video generators still let you create for free—you just need to know which one to use for what.
Kling AI is a great place to start when you want realistic movement and cinematic-looking clips. Hailuo AI is ideal for quick experiments and bringing still images to life. Pika is the fun one, especially for transformations and reality-bending effects like melting, squishing, or changing objects.
Then there's PixVerse, particularly useful for short-form and stylized social content, while Vidu is a handy option when you simply want to visualize an idea before investing more time or money in it.
None of them is perfect. That's also not the point.
Sometimes “good enough, free and finished in five minutes” beats “perfect, expensive and still rendering tomorrow.”
We broke down all five, what each does best, and which one to open depending on what you're trying to create.
From The Archive:
Back in June 2024, we tried TopviewAI, a tool promising to turn product links, images, text, and existing footage into ready-to-post videos.
And we were impressed.
We gave it 4.5 out of 5, praising how quickly it could create social content, suggest hooks, generate scripts, and assemble everything into something that actually looked usable.
At the time, that felt like a pretty big deal.
Fast forward to 2026, and our expectations have changed dramatically.
Today, we're asking AI to generate the actual footage. We expect realistic movement, image-to-video animation, cinematic camera moves, visual effects, character consistency, and increasingly convincing physics. And preferably we'd like to try all of that for free.
What seemed impressive in 2024 hasn't become irrelevant—quite the opposite. TopviewAI was an early glimpse of where video creation was heading: less time spent assembling everything manually and more time deciding what you actually want to make.
It's fascinating to look back and see how quickly the definition of an “AI video tool” has changed.
Two years ago, having AI help make your video was exciting.
Today, we're surprised when it can't make the whole thing.
👉 Read our original review: Video-Making App That Even Gen Z Likes
Text generation | Image Generation | LMAI recommends |
|---|---|---|
Like Magic AI NFT 🏞️
Our master plan is to publish an NFT image in each newsletter and hand it out to our subscribers. The earlier you subscribe, the smaller the series are. It's a future collectible, a piece of digital art that captures the essence of this moment in time.
Thank you for being a valued subscriber. Together, let's embrace the magic of AI and creativity!
LMAI226-21082026
Was this email forwarded to you? Sign up here 👇
