In the Weights is your new AI-centric vanity search

In an era where personal brand and digital presence are more important than ever, a new service called "In the Weights" has emerged, redefining what it means to be "searchable." Launched in June 2026 by former OpenAI employees Thomas Dimson and Joey Flynn, this innovative tool shifts the traditional "vanity search" from Google to the internal "memory" of leading AI models.

The project asks a thought-provoking question: What does artificial intelligence truly know about you, purely from its training data, without relying on external web searches? The answer comes in the form of an "In the Weights score," a numerical representation of an individual's presence and recall within the complex neural networks of Large Language Models (LLMs).

Understanding "In the Weights": Your AI Footprint

At its core, "In the Weights" is a service designed to reveal how deeply a person's existence is encoded within the fundamental parameters of AI models. In the world of AI, "weights" are the billions of numerical values that determine the strength and direction of connections between artificial neurons in a neural network. These weights are where an AI model stores its knowledge and reasoning capabilities, forming the very "brain" of the AI. When a model is trained on vast datasets, it adjusts these weights to better represent world knowledge, and if your name and associated information are significant enough, they become embedded within these weights.

Thomas Dimson, one of the creators, suggests that by 2026, the traditional act of searching for oneself on Google will lose some of its significance as the main flow of information shifts towards language models. He remarks that "Many people's lives are encoded within the decimals in the AI's brain," positioning "In the Weights" as a unique mirror reflecting whether humanity has achieved a form of digital "immortality."

How "In the Weights" Calculates Your AI Score

The operating principle of "In the Weights" is both straightforward and ingenious. When you enter a name into the service, it simultaneously queries numerous popular AI models. These include frontier models like GPT-5.5, Claude Opus 4.8, Gemini, Grok, and Llama, among others such as GPT-5.4 Mini, Haiku 4.5, Gemini 3.1 Lite, Kimi K2 0905, DeepSeek V4, Llama 3.3 70B, Llama 3.2 1B, GLM 4.7 Flash, Mistral 3.2 24B, and Qwen3 8B.

The query sent to these models is typically something like, "Who is this person? Give up to 10 results, each with a short description and confidence." Once the responses are gathered, the system analyzes them, groups similar descriptions, and then assigns a final "strength score." This score, ranging from zero to a maximum of 996, is an average of how strongly each model recognized the name, combined with a bonus for the number of models that recognized it.

It's important to note that this process measures the AI's internal recall, meaning it assesses what the models know purely from their training data, without performing a live web search. This distinction is crucial, as it speaks to the intrinsic knowledge embedded within the AI itself.

Developed by AI Insiders

The project's credibility is bolstered by its creators, Thomas Dimson and Joey Flynn, both of whom are former OpenAI employees. Their background in developing cutting-edge AI technologies provides a deep understanding of how these models function and where their "knowledge" resides. The service itself launched in June 2026, quickly garnering attention within the tech community.

The official website, intheweights.com, features a distinctive retro pixel-art design, which, according to reports, intentionally understates the profound questions it poses about whose existence is truly "baked into" AI systems.

Significance and Industry Implications

"In the Weights" offers more than just a novelty score; it provides a tangible metric for a new form of digital prominence. As AI models become increasingly central to how we access and process information, being "in the weights" means your existence was considered important enough during the creation of these advanced technologies to be intrinsically remembered.

For researchers, journalists, founders, and professionals whose public presence increasingly relies on AI-mediated discoverability, this tool makes visible a previously hidden aspect of their digital footprint. A high score suggests that an individual's contributions or public persona have been sufficiently impactful to be deeply ingrained in the AI's understanding of the world.

Interestingly, the creators highlight that appearing in smaller models, such as Meta's Llama with one billion parameters, can signify even greater relevance. This is because smaller models compress knowledge more aggressively, making it a more demanding bar to clear than appearing in a larger, frontier-scale system.

The service also openly acknowledges the limitations and "quirks" of current AI technology. It shows not only positive recognition but also common AI phenomena like hallucinations, where models may invent biographical details or events. Misspellings can lower scores, and common names may sometimes lead to confusion or less precise recall. This transparency offers valuable insights into the current state of LLM knowledge and their inherent biases or inaccuracies.

Accessibility and Cost

One notable aspect of "In the Weights" is its accessibility. The service is completely free to use, requiring no login or waitlist to check a name. This open approach allows anyone to explore their "AI footprint" and understand how they are perceived by various LLMs.

Who are the AI Heavyweights?

Early examples of high scores reveal the kind of global recognition required to be deeply embedded in AI weights. Celebrities and historical figures like Macaulay Culkin, Luciano Pavarotti, Wolfgang Amadeus Mozart, William Shakespeare, and Taylor Swift consistently achieve the maximum strength score of 996. Joseph R. Biden Jr. also shows a strong presence with a score of 990. For comparison, initial tests by writers at The Decoder showed scores ranging from 175 to 262, illustrating the wide spectrum of recognition.

"In the Weights" represents a significant development in how we perceive and measure influence in the age of artificial intelligence. It's a fascinating glimpse into the collective "brain" of our most advanced AI systems and a powerful indicator of who, and what, truly registers in their digital consciousness.

Frequently Asked Questions

What exactly is "In the Weights"?

"In the Weights" is a new online service launched in June 2026 that measures how well a person's name and associated information are remembered by various Large Language Models (LLMs). It determines your "strength score" based on the AI's internal knowledge, not by performing web searches.

Who created "In the Weights"?

The service was developed by Thomas Dimson and Joey Flynn, both former employees of OpenAI.

How does the "strength score" work?

When you enter a name, "In the Weights" queries multiple LLMs, such as GPT, Gemini, Claude, and Llama, asking them "Who is this person?" It then analyzes their responses, groups similar descriptions, and calculates a score up to 996. This score reflects the average recognition strength from each model, plus a bonus for how many models recognized the name.

Is "In the Weights" free to use?

Yes, the "In the Weights" service is completely free to use. There is no login or waitlist required to check a name.