Hi fam!
We're still recovering from celebrating our 200th newsletter 🥳 (and yes, we're still slightly amazed we made it this far).
While the confetti settles, we've been thinking a lot about how AI is quietly changing the way we work. Not with giant breakthroughs every day, but through small habits, tiny automations and workflows that save a few minutes here and there — until suddenly you're saving hours.
So lately we've been testing new tools, revisiting old favorites, and looking for those practical AI shortcuts that make work feel just a little more like magic. ✨
The future of AI feels Like Magic, and it’s here!
Here are our top AI stories worth your attention today.
Because keeping up with AI shouldn't require opening 37 browser tabs. 😄
A Hugging Face team member shared a reboot of PapersWithCode-style infrastructure at paperswithcode.co, with AI-assisted parsing of papers, leaderboards, and methods. Why it matters: Research discovery and benchmark tracking are core AI infrastructure, and the community has badly wanted a real successor.
Linus Torvalds says AI bug reports are becoming “almost entirely unmanageable”. Torvalds warned that duplicate, AI-generated security reports are wasting maintainer time, especially when sent privately. He’s not anti-AI, but wants contributors to use it productively and ideally submit patches, not just noisy reports. Why it matters: This is a real-world example of AI increasing coordination costs even when it improves raw detection.
The Vatican says the Pope’s first encyclical, Magnifica humanitas, will address preserving the human person in the age of AI. Anthropic co-founder Christopher Olah is listed among the event speakers. Why it matters: AI governance is now fully crossing into major religious and moral institutions, not just tech and regulation circles.
Simon Willison’s “last six months in LLMs in five minutes” became the top AI discussion on HN. The piece argues that the big shift was not just better models, but coding agents becoming good enough for daily use. It also highlights the rapid improvement of open-weight and laptop-runnable models. Why it matters: Not breaking news, but it was the most influential “state of the field” discussion item today.
Andon Labs’ AI-run radio station experiment is getting a lot of attention. Andon let four different models run autonomous radio stations over months, handling programming, music choices, replies, and scheduling. The result was a mix of impressive persistence and weird behavioral drift. Why it matters: It’s a useful real-world look at what long-running autonomous agents actually do when left operating continuously.
🎨📷 Reverse Prompt Hack
We’ve written a lot about generating AI images over the years.
Lighting. Composition. Camera angles. Art styles. Color palettes. Aspect ratios.
But today, let’s flip the script.
Instead of asking:
“How do I write a good image prompt?”
Ask:
“What prompt created this image?”
Simply upload an image you like and use this prompt:
Analyze this image and write the prompt that most likely generated it. Include the artistic style, composition, lighting, camera angle, color palette, mood, subject details and any technical parameters you can infer.
The AI will break down the image and identify the ingredients that make it work — from cinematic lighting and lens choices to styling, colors and composition.
It's a fantastic way to learn prompt writing because you're studying examples that already produced a result you love.
Instead of starting with a blank page and guessing what to type, you reverse-engineer successful images and discover the patterns behind them.
After a few examples, you'll start noticing how often great AI visuals rely on the same building blocks — and writing strong prompts suddenly becomes much easier. ✨
The Fastest Way to Get Better Answers From AI
Most people still use AI the same way they did two years ago:
Open AI assistant.
Type a question.
Wait for an answer.
But one of the biggest AI productivity hacks has nothing to do with writing better prompts.
It’s using your camera instead of your keyboard. 📱
Need help understanding a spreadsheet? Take a screenshot.
Got an error message? Upload it.
Looking at a PDF, report or handwritten notes? Show it to AI instead of spending five minutes explaining what you're looking at.
Modern AI tools can understand images, documents, charts and screenshots surprisingly well — and often give better answers because they have more context.
In the full article, we share 5 practical examples where showing AI something is faster (and smarter) than typing it out yourself.
Read the full article here 👉 The Fastest Way to Get Better Answers From AI
From the Archive: 🎬 How to create a professional explainer video from text
Long before AI video generators became mainstream, we tested a simple idea:
Can you create a professional explainer video starting with nothing but text?
The answer was yes — and the workflow has only become more powerful since then.
If you're creating tutorials, educational content, product demos or social media videos, this article is a great reminder of how quickly an idea can turn into a finished video with the right AI tools.
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!
LMAI201-22052026
Was this email forwarded to you? Sign up here 👇
