Sora's shutdown is pushing creators everywhere else 🏃

Good morning. Sora's final shutdown date is locked, and creators are scrambling onto other tools before the lights go out. A new platform will take a song and spit out a fully storyboarded music video on its own. And Facebook is flooding your feed with AI "silver fox" podcast bros dispensing relationship advice to no one. Which of these is hitting your feed hardest? Hit reply and tell me. We cover all three below.

Sora's API goes dark for good on September 24, the last piece of OpenAI's video tool to shut down after the consumer app already closed back in April. Anyone who built a workflow on Sora has a hard deadline to get off it, and the exit is well underway. Platforms like InVideo and Hedra are actively rerouting people onto Veo 3.1, Kling 3, Seedance 2.0, and MiniMax H3 for the synced-audio, multi-shot work Sora used to handle.

OpenAI had the most hyped video model on earth barely a year ago and is walking away from it. Video generation burns cash at a rate text never does, the app bled users, and OpenAI would rather aim that compute at enterprise. The flashiest tool in the space just became a cautionary tale about building your whole setup on one company's model.

Don't marry a single generator. The people who saw Sora coming apart and had already spread their work across Veo, Kling, and Seedance barely felt it. The people who went all in on one tool are now rebuilding under a countdown. Keep your process portable and learn more than one model.

A new platform called OiiOii AI launched a feature that takes a full song and builds a complete music video around it, not one clip, the whole thing. It listens to the track, reads the lyrics, picks up on the rhythm and mood, then writes a script, storyboards it, generates the characters and scenes, adds camera movement, and edits it with subtitles and color. Three modes: narrative for story videos, abstract for dreamy visual stuff, performance for artist-focused promos.

OiiOii says it spits out a polished first cut in about three minutes, which is a company claim, not a tested number, so treat it as a starting point you'll still refine. The pitch is that it interprets the song instead of just generating clips off a prompt, which is the actual hard part of a music video, matching the visuals to the beats and the emotional turns.

It might not be good enough yet, but tools are clearly moving from "make me a clip" to "make me the finished piece." For a musician who can't afford a director and an editor, one that turns a track into a watchable video by itself is a real unlock, even if the first cut needs work. Worth a test the next time you or someone you know drops a song.

Facebook's algorithm spent the past week shoving a weird new genre at everyone: hyper-realistic older, handsome men sitting at podcast mics, talking straight to camera, dispensing relationship advice. Think "Uncle George," a silver-haired gent in a button-down giving "real talk on love and loyalty" to women tired of confusion. None of it is real. Not the man, not the podcast, not the show it's supposedly clipped from.

The whole thing is built to look like a clip pulled from a real podcast, which is exactly why it works, that format reads as authentic. The giveaway is the script, stuffed with "it's not this, it's that" phrasing, the same robotic cadence that outs AI writing everywhere. It's engagement bait, engineered to farm comments and shares from people who don't clock that the guy doesn't exist. The fake-podcast-clip format is clearly catnip for the algorithm right now, so expect a wave of copycats.

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