Good morning. Alibaba dropped a video model that makes 30-second clips with sound and can turn a slide deck straight into a video. California's new transparency law just kicked in and it's about to force AI watermarks onto most of the tools you use. And half of Ireland spent the weekend convinced the Cat in the Hat was prowling their streets at night, thanks to AI. Which of these are you clicking first? Hit reply and tell me. We cover all three below.
📄 Alibaba's New Video Model Turns a PDF Into a Video
Alibaba rolled out the full launch of Wan 3.0 this week, its newest video model, after a quiet beta earlier in the month. It makes clips up to 30 seconds long at solid 1080p quality with sound generated right in, all in one shot instead of stitching short pieces together.
Their newest feature: hand it a PDF, a spreadsheet, or a slide deck and it reads what's inside and builds a video out of it. A pitch deck becomes a promo, a report becomes an explainer, without you scripting and storyboarding from a blank page. For anyone who makes faceless content or turns written stuff into video, that collapses a whole workflow into one upload.
It currently runs through Alibaba's cloud, and you pay per second of video. Still, native audio plus fixed faces plus document input in one model is a real step. Worth testing on your next video idea.
California's New Law Is About to Watermark Half Your Tools
California's AI Transparency Act went live on August 2, the same week Europe's version did, and it hits any big AI platform with more than a million users in the state. Those companies now have to invisibly watermark the image, video, and audio their tools generate, offer a free tool anyone can use to check whether something was AI-made, and give you the option to put a visible AI label on it. Skip it and they're looking at fines that stack up fast, five grand a pop.
This lands on the tool-makers, not you directly, but it flows straight downhill. The AI video, image, and voice tools you use are the "covered platforms," so their output is about to carry a hidden AI marker. Pair that with Claude's watermarking from last week and the direction is clear: an invisible "made by AI" tag is becoming standard equipment across the whole stack.
One thing to keep straight: this covers image, video, and audio, not written text, and it only forces the big platforms, not your personal workflow. Nothing here says you can't use AI. It says the tools have to leave a detectable fingerprint, and anyone with the free checker can read it.
Half of Ireland Thought the Cat in the Hat Was Stalking Them
Over the weekend a genuinely unhinged AI panic swept the UK and Ireland: fake images and videos of the Cat in the Hat, styled after the 2003 Mike Myers movie, crouching on parked cars and peering into people's windows at night. It spread far enough that actual police forces had to post statements telling people the six-foot cartoon cat was not, in fact, hiding in their driveway.
It started in Limerick and jumped across counties and then the whole UK, with residents genuinely scared and one clip pulling 15 million views in a few weeks. The format was simple, drop a creepy familiar character into ordinary night-time footage of a normal neighborhood, which is exactly the kind of thing these tools now nail. Familiar face plus a boring real-looking street plus dark equals believable enough to freak people out. Could this be the trend this Halloween?
Discussion
Stuck on a tool or have an AI video question? Reply, I’ll do my best to help.
Created something cool? Send it over and I might feature you. 📺
📫 Want to get your tool featured or advertise to our readers? Let's chat → [email protected]
Thanks for tuning in!
-Kevin