Meta released Muse Image this week. The company now competes with OpenAI and Midjourney and Stability AI in a market where the primary business model is letting people generate anime girls with improbable proportions. Zuckerberg wants advertisers to use it. He wants creators to subscribe to it. He wants you to believe this changes everything about how brands will reach consumers in 2026.
The pitch goes like this. Advertisers pay Meta to generate images instead of paying actual photographers or illustrators. Creators pay Meta to generate images instead of learning to use a camera. Everyone saves money except the people who used to make images for a living. Meta calls this innovation. The rest of us call it a race to replace human labor with statistical pattern matching that occasionally gives people six fingers.
Here's what actually happens. Some guy who runs a dropshipping store selling resistance bands generates forty-seven product images of a man with abs that defy human anatomy. He split-tests them all. Conversion rate stays exactly the same. He blames the algorithm. He generates forty-seven more images. Facebook shows them to nobody because his ad account got flagged for using AI slop. He goes to Reddit and complains about how Meta is rigged against small businesses. The cycle continues.
Meta enters a race that has no finish line and no winner. Every AI lab pumps out an image model. Every model generates pictures that look vaguely wrong in ways you can't articulate until you notice the earrings phase through the neck. Every company promises this will revolutionize advertising. Every advertiser still just wants to know why their cost per click went up thirty percent last quarter.
The real breakthrough will come when Meta builds an AI that can explain why anyone thought training a model on billions of images so brands can generate stock photos of people holding coffee was worth the energy consumption of a small nation.
Photo by Julio Lopez on Unsplash
Leave a Comment