In the Weights is your new AI-centric vanity search

Measuring Notability in the Age of AI

Two former OpenAI employees, Thomas Dimson and Joey Flynn, have launched a new tool called In the Weights, which quantifies an individual's prominence within the training data of various large language models. This project offers a modern alternative to the traditional Google vanity search, reflecting a growing awareness that AI models, not just web crawlers, are becoming a primary source for public information. The site's leaderboard-style scoring system taps into a desire to see if one's existence has been deemed significant enough to be encoded into the foundational knowledge of AI.

How 'In the Weights' Operates

The platform functions by querying multiple prominent LLMs to gauge how well they can recall information about a specific person without accessing the live web. It then aggregates and scores the outputs to generate a 'strength score' indicating how deeply a person is embedded in the models' collective knowledge. The founders' goal was to explore the idea that 'so many lives are encoded somehow in a bunch of floating point numbers inside the AI brain'.

• Models Queried: The system queries a diverse set of models including Grok, Gemini, multiple GPT versions, Claude, and Llama.
• Process: It prompts models with a question like 'Who is <name>?', then clusters similar descriptions from the generated responses.
• Scoring: A proprietary 'strength score' is assigned based on the consistency and detail of the recalled information across all models.
• Features: The results highlight potential hallucinations and specify which models provided which answers, offering a comparative look at their internal knowledge bases.

The New Metric for Digital Permanence

While framed as a lighthearted tool, In the Weights points to a significant cultural shift where being 'encoded' in an AI's parameters is a new form of digital relevance. The platform also serves as an informal tool for analyzing the inherent biases and knowledge gaps between different AI systems. This could foreshadow a future where an individual's or entity's influence is measured not just by search engine ranking, but by their persistent presence within the core training data of foundational AI models.

Strategic Takeaway: The emergence of tools like In the Weights indicates that the battle for relevance is shifting from search engine optimization to 'model optimization.' As AI becomes a primary information interface, presence within model weights will be the new metric for digital notability, influencing public perception and creating a new class of data analysis focused on model bias and knowledge gaps.