The premiere episode of the AIography Podcast is up. I sat down with Robert Gaudette. He's a 54-year-old nonprofit worker from Toronto. He never went to film school. And he just won the top prize at Runway's AI Grand Prix with his film "A Face Only a Mother Could Love." Here's what got me. The film has real heart. It moves people. That's the one thing everyone says AI can't do, and Robert did it anyway. We talked through how he actually built it, start to finish. His workflow. The "subtext block" he puts in his prompts. How he pushed past the imposter feeling that hits all of us. And why he still believes story comes first, even now. If you've been stuck at "where do I even begin," this is the one to watch. Robert started exactly where you are. Link to Interview Link to Robert's film "A Face Only A Mother Could Love." Watch it, then drop one thing that surprised you in the comments.

If you've felt whiplash this month, you're not imagining it. One week a festival puts a fully AI-made film in its lineup. The next, a craft guild publicly tears into a legendary director for promoting an AI tool. And now the Directors Guild has a tentative deal that writes AI rules into the contract. The signals point in every direction at once. Here's the thing I want you to hear, because nobody seems to be saying it plainly. The split isn't tech people versus artists. It runs straight through the working pros. People doing the exact same job are landing on opposite sides. The colorist feels it at a different moment than the location scout. Both reactions are completely rational. What I've noticed across a long career of these turns is that the divide tracks one thing: exposure. Whoever's paid work the tool can already touch feels it first and feels it hardest. That's not fear talking. That's people doing math about their own livelihood, and they're not wrong to. So if you're standing in the middle of this, confused about which side you're supposed to be on, here's my honest read. You don't have to pick a tribe. The belief-system version of this argument is a trap. The useful version is a much smaller, calmer question. Try this: Make two short lists. On the left are the parts of your work the tool can genuinely touch today. On the right are the parts it can't. Be honest, not hopeful, on both sides. Then get fluent in the things on the left before the decision gets made for you, and stop losing sleep over the things on the right. The people who do worst in a turn like this treat it as a fight to win. The people who do best treat it as a skill to learn. You don't have to love it or hate it. You have to know exactly where it lands in your work and act from there. Where do you land right now, and which part of your job made you land there? I read every reply. (Founding Members: this week I'm breaking down the single hardest technical problem in AI video right now, keeping a character looking the same from shot to shot, and the workflow tricks that actually help today.)