In the Weights is your new AI-centric vanity search
Original reporting by TechCrunch
In the Weights is a novel web service designed to quantify how deeply an individual's identity is embedded within the 'weights'—the numerical parameters—of various large language models. Born from a recognition that traditional web search is no longer the sole arbiter of public information, with chatbots increasingly becoming primary sources, creators Thomas Dimson and Joey Flynn sought to measure how well AI models recall someone without external tools. The service queries a diverse array of models, including Grok, Gemini, multiple GPT versions, Claude, and Llama, asking, "Who is <name>?" It then clusters descriptions and assigns a "strength score," indicating an individual's digital footprint within the AI's "brain."
Beyond the search bar
This innovative approach goes beyond traditional online vanity searches. Dimson and Flynn, both formerly of OpenAI, conceived In the Weights to explore the idea that "so many lives are encoded somehow in a bunch of floating point numbers inside the AI brain." The site has quickly resonated, generating "insane" reception by tapping into a collective curiosity about digital permanence. Users can now see if they "live forever in the super intelligence," with a dynamic leaderboard showcasing everyone from everyday tech enthusiasts (like this article's author, who scored 641) to global celebrities like Macaulay Culkin (currently with a top score of 988). Beyond the intriguing comparisons, In the Weights also illuminates potential AI hallucinations, such as GPT-5.4 Mini's ambiguity regarding common names. This unique lens not only sparks playful competition but also prompts deeper questions about AI's internal biases and its evolving role in shaping collective memory.
In the Weights emerges as more than just a playful curiosity; it is a resonant barometer of our evolving digital existence. By quantifying an individual's presence within the core parameters of leading AI models, the platform starkly illustrates a pivotal shift in how information about people is not just retrieved, but inherently stored and understood. This novel metric introduces a new form of "vanity search" for the AI age, challenging the long-held supremacy of traditional web search as the sole arbiter of notability and introducing a deeper, perhaps more profound, dimension to our digital footprint.
Redefining Notability
The enthusiastic reception for In the Weights underscores a fundamental human desire for recognition, now recontextualized for an AI-first world. This paradigm shift carries significant implications for how reputation is built and managed, how personal information is disseminated, and even the very concept of individual identity in an increasingly algorithm-driven information ecosystem. As AI models become primary conduits for knowledge, their internal biases, recall capabilities, and even their "hallucinations" will inevitably shape public perception in ways traditional search engines never could. Future insights from platforms like In the Weights, exploring model disparities, biases towards certain demographics, and overlooked figures, will be crucial. They offer a unique lens into how these powerful systems are currently encoding human experience, hinting at a future where our digital legacy is as much a function of an AI's internal "weights" as it is of traditional media and human memory.
Frequently asked questions
- What is In the Weights, and how does it assess AI models' knowledge about individuals?
- In the Weights is a website that measures how well large language models (LLMs) can recall information about specific individuals without relying on real-time web search. It queries various AI models with names and assigns a "strength score" based on the quantity and consistency of the information returned, reflecting how "ingrained" a person's data is within the model's internal parameters or "weights."
- How does the In the Weights platform calculate a "strength score" for individuals in AI systems?
- The platform queries multiple large language models (LLMs) with an individual's name, requesting descriptions and a confidence score. It then clusters similar descriptions and assigns a numerical "strength score." This score reflects the breadth and consistency of the information the AI models can recall from their pre-trained data, indicating an individual's presence within the models' internal "weights."
- Why are individuals interested in having their information encoded within artificial intelligence models?
- People are interested in their information being "in the weights" of AI models because it suggests a level of digital permanence and significance within the rapidly evolving AI landscape. As more information consumption shifts to large language models, being "remembered" by AI is seen by some as a new form of recognition, a modern vanity search, or even a pathway to digital immortality.