In the Weights: A New Kind of Vanity Search for the LLM Era

In the Weights Launches as an AI-Native Vanity Metric

A new website called In the Weights is letting people check how well AI models remember them without relying on traditional web search. According to TechCrunch, the site was created by Thomas Dimson and Joey Flynn, who both previously worked at OpenAI after it acquired their design startup Global Illumination. The tool queries multiple large language models—including Grok, Gemini, various GPT versions, Claude, and Llama—asking each to recall information about a given person, then clusters the responses and assigns a strength score based on model consensus.

The “weights” in the site’s name refer to the numerical parameters that define how an AI model behaves and what it has learned. Dimson and Flynn’s premise is that being encoded into those parameters—having your existence recognized by multiple AI systems—represents a new form of cultural presence in the age of language models, as traditional Google vanity searches lose relevance.

How the Ranking System Works

In the Weights returns a strength score for each query, along with individual model responses and potential hallucinations. According to TechCrunch’s reporting, the site highlighted a case where GPT-5.4 Mini misidentified a person’s name as “an ambiguous name form that could refer to multiple people with the initials A.H.A.” The leaderboard includes celebrity examples: actor Macaulay Culkin currently ranks at the top with a score of 988, followed closely by opera singer Luciano Pavarotti.

The strength scores are dynamic. TechCrunch noted that the leaderboard shifted measurably during the time it took to write the article, suggesting the rankings adjust as models are queried or as model versions update. The site also surfaces which specific models returned which answers, allowing users to see where disagreement occurs.

Why Creators Built It and Reception

Dimson told TechCrunch that he and Flynn were seeking a creative project after leaving OpenAI, and the idea crystallized around the observation that “Google vanity searches are the wrong objective in 2026 as more traffic moves to LLMs.” He also cited curiosity about how individual identities are “encoded somehow in a bunch of floating point numbers inside the AI brain,” and drew inspiration from a playful blog post comparing AI weights to Terry Bisson’s short story “They’re Made Out of Meat.”

According to TechCrunch, Dimson said the reception had exceeded expectations: “Reception has been insane so far, we thought this would be a mild curiosity but it seems like it has struck a nerve of wanting to see if you live forever in the super intelligence.” The comparison-based leaderboard mechanic appears to have amplified engagement by tapping into existing vanity-search behavior while reframing it around AI model training.

Why This Matters

In the Weights reflects a meaningful cultural shift: as people increasingly encounter information through chatbot interfaces rather than search engines, the question of who or what gets encoded into AI training data takes on new social weight. The site is essentially a public mirror for AI model training patterns, allowing anyone to discover their position (or lack thereof) in the “attention landscape” of multiple LLMs at once.

For AI researchers and practitioners, the tool illustrates how opaque model training truly is—the hallucinations and inconsistencies TechCrunch documented highlight real gaps in model knowledge and potential biases. For users, it dramatizes the shift from “Did Google find me?” to “Did the AI learn about me?” as the meaningful measure of public presence. Whether this is genuinely consequential or playful cultural commentary remains an open question, but the rapid user adoption suggests the framing has resonated.