Jevtown
Town of 10,000 computed personas where Jev scores a post, listing, product, or headline against about 60 audience attributes to plan who sees it, then answers one batched Choice…
# Jevtown Live at [jevtown.ivanhabor.com](https://jevtown.ivanhabor.com), no sign-in. [Watch the 30-second video](https://www.youtube.com/watch?v=Ktm2qwW7JAo). [](https://jevtown.ivanhabor.com) A social network where people write and 10,000 AI personas read. Post a text, a listing, a product or a headline, and the crowd reacts within seconds: most scroll past, some like, repost, block, write to the seller or buy. Every reaction comes from [Jev](https://typesafe.ai), a model that answers typed questions with probabilities and writes no text. What you get for a text: - **The map of its audience.** 100×100 dots, one per persona, neighbours are similar people. A text starts with the 600 people Jev thinks it is for, and travels further only while the crowd is glad to see it. A good post visibly spreads across the map; spam, rage and scam listings die in the first wave. Every post in the feed carries its map, and the big map on the right always shows the crowd of the post you are looking at. - **Who stopped, who liked it, who got annoyed**, by interest, job, age, temper, budget, city and what people are shopping for. Point at a group and its people light up on the map. - **Voices of the crowd.** The people who reacted, by name: "Svitlana, 62, civil servant, Kharkiv · wrote to the seller · «Can I check it before paying?»". - **For a listing:** what buyers would ask the seller first, out of twenty prepared questions. - **For a product:** how many would buy at each price of a ladder you set, and which price earns the most. - **The result in one sentence**: how far the text went, which wave stopped it, how many were glad and how many annoyed. - **Versions.** Edit the text, show it again, see what changed, with the two texts side by side. - **Your resident** (`/me`). Make up