AI
In the Weights: A New AI Vanity Search That Measures Your Digital Immortality
Ex-OpenAI designers created a tool that ranks how well AI models remember you without web search—revealing whose existence is "encoded" into AI brains.
TechCrunch — AI
Google vanity searches might be obsolete. With chatbots increasingly replacing traditional search as people's go-to information source, Thomas Dimson and Joey Flynn—both former OpenAI designers—built a different kind of ego check: In the Weights.
The tool queries multiple AI models (GPT, Claude, Gemini, Grok, Llama, and others) with the same question: "Who is [name]?" It then aggregates their responses, clusters similar answers, and assigns a "strength score" reflecting how embedded you are in each model's training data.
The Leaderboard Effect#
The results are presented as a ranked leaderboard. Macaulay Culkin and opera singer Luciano Pavarotti currently top the chart with strength scores near 1,000. Lesser-known figures rank lower—demonstrating an obvious bias toward celebrities and public figures. The site also flags potential hallucinations, such as GPT-5.4 Mini confusing "Anthony Ha" with an ambiguous acronym.
Dimson told TechCrunch the project emerged from thinking about how "Google vanity searches are the wrong objective in 2026 as more traffic moves to LLMs," and about the philosophical weight of human identity encoded in AI parameters. The creators were inspired by a tongue-in-cheek riff on AI weights and Terry Bisson's sci-fi short story "They're Made out of Meat."
What It Reveals#
The tool's reception has been unexpectedly strong, resonating with people curious about whether they've achieved digital permanence in AI systems. It also exposes disparities: different models in the same series return wildly different answers about the same person, and it highlights whose stories are captured by AI training versus whose are missing.
Dimson plans to investigate deeper—analyzing model biases, discrepancies between versions, and which people "should have a Wikipedia article but don't."

