The Big Story

A company called Icon once raised millions to build an AI tool that would turn any website link into dozens of short video ads using fake people on screen. The idea sounded perfect for social media: cheap, fast, and endless variations. But after the site went quiet for a while, the team came back with a completely different product called "The Human Admaker." Instead of generating videos from code, they now hire real creators, write scripts with them, coach the filming, and deliver finished ads. The surprising part is the results. Even though human-made videos cost more upfront, they turn into sales far more often because viewers trust them. Think about your own feed — you've probably scrolled past dozens of stiff, too-perfect clips that feel like they came from a machine. This change shows that for some jobs, AI works best when a person stays in charge of the final step. For you, it means the ads you see might start feeling more real again, and it hints at future careers where knowing how to guide AI (instead of letting it run alone) becomes a valuable skill. If you make videos for school projects or a small online shop, this story is a reminder that mixing your own ideas with AI tools often beats pure automation. You can try this idea yourself by taking a short video on your phone and editing it with a free app like CapCut, then compare how it feels next to something made entirely by an AI image or video tool. Source: x.com

Explain Like I'm 14

You know how you sort your stuff into groups without even thinking — like putting your phone in the "electronics" pile and your hoodie in the "clothes" pile? Now imagine you do the same thing with big ideas like "learning." Most people picture learning as something slow, curious, and a little messy, like a kid figuring out why a bike stays up. When they hear the word AI, their brain quickly puts it in a different pile labeled "machine, not alive, just code." That quick sorting happens before they even look at what the AI actually does. So even when an AI spots a pattern in new situations and uses it to answer something it never saw before, people still say it isn't learning — because the label was already decided. The tricky part is that the same person can give a definition of learning that perfectly matches what the AI is doing, yet still refuse to call it learning if you ask the questions in the wrong order. It's like the brain protects the category it already chose. The honest spot in the middle is admitting you don't know exactly what's happening inside the system, but noticing that treating it like it has some kind of thought changes how it responds. That middle spot feels uncomfortable because schools and companies like clear yes-or-no answers, but it might be the most accurate way to look at these tools right now. Source: reddit.com

Cool Stuff & Try This

A phone screen with literally zero borders TECNO just showed a concept phone where the display goes all the way to the very edge on every side — no notch, no thin line, nothing. It looks like the picture is floating in your hand with no frame around it. This matters because phone screens keep getting bigger while the body stays the same size, and removing the last bit of border changes how apps and videos feel. If you like photography or watching shows on your phone, a screen like this would make everything feel more immersive. You can see what it looks like right now by searching "TECNO 0mm Bezel Concept Phone" on YouTube and watching the short clip from Marques Brownlee — it only takes a minute and shows the phone from every angle. Source: x.com

Asking AI about its own environmental footprint Google's Gemini gave clear answers when someone asked how data centers could use less electricity and water while still running chatbots. The suggestions included building centers in cold places so outside air does the cooling for free, switching to sealed liquid systems that never evaporate water, and pairing each center with its own solar or wind power so it doesn't pull from the local grid. These ideas connect directly to your daily life because every time you use a chatbot, somewhere a server is working and creating heat. You can try this yourself by opening Gemini (or any free chatbot) and asking it the same question about data center resources — see what solutions it suggests and compare them to what you already know about energy use. Source: reddit.com

Quick Bits

Robots racing until they burst into flames At the world humanoid robot games, some machines ran so fast they set new speed records — and a few literally caught fire from the effort. The events that looked more like household chores (folding laundry, carrying boxes) turned out to be harder for the robots than the flashy races. It's a fun reminder that making robots helpful in real homes is still a bigger challenge than making them go fast. Source: arstechnica.com

ChatGPT stepped in to stop a crime In one recent case, ChatGPT helped prevent a planned attack by giving advice that led to the person being stopped before anyone got hurt. Stories like this show the same tools that help with homework can also be used in serious, unexpected ways when someone reaches out for help. Source: open.kg

```claims []

💬 Reply to this email — Patrick reads every one.

Share: X · LinkedIn · WhatsApp

Forwarded this email? Subscribe here — it's free.