Escaping the Matrix: Testing Your Fame Against AI Memory Banks
If you Google your name lately, you probably notice it does not feel the same. Standard web search is losing its crown as the main source of online truth. Instead, millions of users are turning to AI chatbots to figure out who people are. This major cultural shift led builders Thomas Dimson and Joey Flynn to build a fascinating new tool called In the Weights. This platform treats your online reputation like an entirely new game by measuring exactly how well a machine learning model remembers you without using actual web search tools.
The mechanics of the site depend on the core settings of large language models. The weights mentioned in the project name represent the actual mathematical values that control an artificial intelligence model’s knowledge base. To see how famous you are to a machine, the system queries various major AI models, including Grok, Gemini, multiple versions of GPT, Claude, and Llama. It asks them a simple question: “Who is this person?” The tool collects up to ten descriptions, groups similar answers, and generates a concrete numerical strength score for your name.
If you score high on the platform, it means a neural network decided your personal data was important enough to keep during its main training phase. For example, tech writers testing the system saw massive variations in their personal rankings. While a standard blogger might pull in a modest strength score that lands them in the top six percent of global searches, notable internet figures and journalists achieve much higher scores. The live leaderboard shows famous actors and singers holding the top slots with massive strength numbers, proving exactly who holds a permanent spot inside the AI memory bank.
The tool also uncovers strange bugs and biases within current AI structures. When you test a name, the results show which specific models suffer from severe hallucinations or confusing mix-ups. A smaller model might get confused by common names, blending separate real-world people into a single fictional profile. This reveals a massive flaw in how modern technology tracks human history.
The creators built this tool out of personal curiosity after leaving OpenAI. Dimson and Flynn previously sold their design startup, Global Illumination, to the AI giant before moving on to independent projects. Dimson explained that standard Google vanity searches feel completely outdated now that online traffic is shifting heavily toward conversational chatbots. He wanted to show people that their life history is now encoded inside a massive cloud of floating point numbers.
Public reception to the launch has been intense. While many users treat it as a fun internet game, the platform touches on a deeper human desire to see if your legacy will survive inside future super intelligence systems. Critics point out that looking at these lists is basically like forcing a dozen separate digital assistants to gossip about your life behind your back. The layout plays into this feeling by using a charming, retro interface that looks like an old console game. Dimson intends to expand the system to analyze why different versions of the same software produce wildly different summaries and which specific groups of people the data biases leave out entirely.
