Key Takeaways
- In the Weights is a viral new tool designed for the ultimate AI vanity search, probing whether LLMs remember you without using web search tools.
- Developed by former OpenAI engineers, the platform calculates a custom strength score based on direct neural network recall.
- This shift marks a transition from traditional SEO to LLM optimization, changing how creators, artists, and tech professionals track their digital legacy.
For decades, Googling yourself was the ultimate way to gauge your digital footprint, but a new tool called “In the Weights” introduces a fascinating AI vanity search for the modern era. If you have ever wondered whether your name is hardcoded into the neural networks of the world’s most powerful artificial intelligence models, this viral web application provides the answer.
Created by tech innovators Thomas Dimson and Joey Flynn, the project bypasses real-time search engines entirely. Instead, it tests the actual memory banks of today’s most prominent large language models (LLMs) to see if you have achieved digital immortality.
Why the AI Vanity Search is Replacing Traditional Googling
Traditional search engines do not hit the way they used to. With the rise of AI chatbots, web search is no longer the sole canonical source of information about individuals, brands, or creative works.
More people are learning about public figures, artists, and professionals directly from conversational interfaces. Because of this, performing a routine AI vanity search is becoming crucial for understanding how artificial intelligence perceives your identity.
Dimson and Flynn realized this shift after leaving OpenAI, which they joined after their design startup, Global Illumination, was acquired. They wanted to create something that explored how human lives are encoded as floating-point numbers inside the “AI brain.”
How ‘In the Weights’ Measures Your Digital Legacy
The “weights” in an AI model are the numerical parameters that shape its training, knowledge base, and output. To perform its unique AI vanity search, the website queries a vast array of models, including Grok, Gemini, multiple versions of GPT, Claude, Llama, and several lesser-known open-source models.
The platform issues a direct prompt to these models, asking: “Who is [name]? Give up to 10 results, each with a short description and confidence.” Because the tool disables live web-browsing tools, the models must rely entirely on their pre-trained weights.
Once the models respond, the system clusters similar descriptions together. It then assigns a unified strength score that indicates how deeply embedded a person’s identity is within the global AI ecosystem.
This technical process mirrors how advanced neural networks categorize complex visual and textual data. For instance, similar algorithmic principles are used in creative fields, such as the use of generative adversarial networks (GANs) in animation and VFX production to generate and refine complex imagery.
The Leaderboard and the Hallucination Hazard
The results generated by this AI vanity search are both fascinating and highly competitive. The site features a live leaderboard showcasing individuals with the highest strength scores across the internet.
Currently, cultural icons like “Home Alone” star Macaulay Culkin and legendary opera singer Luciano Pavarotti sit near the top of the charts with scores hovering around 988. Meanwhile, prominent tech journalists and digital creators are discovering where they rank in the top percentiles of global AI training data.
However, relying on static weights also exposes the limitations of LLMs, particularly the issue of AI hallucinations. Without live search to verify facts, models frequently mix up identities or invent fictional backgrounds for less-famous names.
For example, GPT-5.4 Mini famously hallucinated that certain tech writers were ambiguous entities or referred to entirely different people. This highlights the double-edged sword of LLM memory: being remembered is one thing, but being remembered accurately is another.
Technical Specifications: How the Models Compare
To understand how different AI models process human identities, the platform categorizes the responses from various LLM families. Below is a breakdown of the primary models queried during an AI vanity search on the platform:
| Model Family | Developer | Memory Style | Hallucination Risk |
|---|---|---|---|
| GPT Series | OpenAI | Highly structured, broad general knowledge | Moderate (prone to initial clustering) |
| Claude Series | Anthropic | Nuanced, context-heavy descriptions | Low (tends to admit ignorance) |
| Gemini Series | Strong integration of historical web data | Low to Moderate | |
| Llama Series | Meta | Diverse, open-source training weights | High (highly dependent on model size) |
| Grok Series | xAI | Conversational, real-time bias in training | Moderate |
A Gamified, Retro Interface for the Future of Search
Part of the charm of “In the Weights” is its visual execution. The website features a retro, Nintendo-inspired design that gamifies the concept of checking your AI footprint.
This playful aesthetic makes the daunting reality of algorithmic surveillance feel approachable and fun. According to the creators in an interview with TechCrunch, the reception has been massive, striking a deep cultural nerve.
People are inherently curious to see if they “live forever” inside the superintelligence. The social comparison factor has also driven viral adoption, as users share their strength scores across social media platforms.
Dimson plans to expand the project further. He aims to investigate why different models within the same series produce wildly different results, which models exhibit regional biases, and which notable figures are missing from Wikipedia but still remain deeply embedded in AI weights.
Frequently Asked Questions (FAQs)
What is an AI vanity search?
An AI vanity search is the practice of querying large language models to see what information they have stored about you in their static training weights, rather than searching the live web.
How does ‘In the Weights’ calculate your score?
The tool queries multiple AI models simultaneously without using live web-search integrations. It then clusters the matching descriptions and assigns a strength score based on the confidence and consistency of the models’ answers.
Why do some AI models hallucinate my information?
AI models do not possess real-time awareness; they predict the next most logical word based on patterns in their training data. If your name is relatively rare, the model may generate plausible-sounding but entirely fabricated facts to fill the gaps.
Who created ‘In the Weights’?
The platform was created by Thomas Dimson and Joey Flynn, two prominent designers and engineers who previously worked at OpenAI after their startup, Global Illumination, was acquired by the AI giant.
