Anyone who has searched on Google lately knows that search results aren’t what they used to be. Sure, everything is happening with Google Search itself, but just as many people are learning about you and me from chatbots, and there’s also an inescapable feeling that web search is no longer the legitimate source of information it once was.
Thomas Dimson and Joey Flynn shared similar sentiments, which led them to found In the Weight. The “weights” in question are numerical parameters that shape the training and output of an AI model, so the website claims they measure “how well a model can remember someone without the use of tools like web searches.”
“Participating in the weigh-in means you are considered important in the process of creating superhuman artificial intelligence,” the website says.
To accomplish this, In the Weights will likely query a variety of models (including Grok, Gemini, multiple versions of GPT, Claude, Llama, and even lesser-known models) with questions like “Who are you? Give up to 10 results, each with a short description and confidence.” It then “clusters similar explanations and assigns a strength score.”
For example, this humble tech blogger received a strength score of 641, putting me in the top 6% of names. I was feeling pretty good until I saw multiple colleagues at TechCrunch getting even higher scores. And as I write this, the leaderboard is changing, with Home Alone star Macaulay Culkin currently in the top slot with an intensity score of 988, followed by opera singer Luciano Pavarotti.
The results also show which models returned an answer for a given name, highlighting potential illusions. Apparently GPT-5.4 Mini says that Anthony Ha is an “ambiguous name form that can refer to more than one person with the initials AHA.”
Asked why he built In the Weights, Dimson told TechCrunch in an email that he wanted to “reignite our creativity” after he and Flynn left OpenAI (they joined the company through its acquisition of design startup Global Illumination).
Dimson said he was thinking about the fact that “Google’s vanity search is the wrong purpose in 2026 as more traffic moves to LLM” and the fact that “so many lives are somehow encoded into a bunch of floating point numbers in an AI’s brain.” He also said the site’s direction was “sealed” by a sarcastic blog post that poked fun at the weight of AI and Terry Bisson’s classic short story “Made of Flesh.”
“The reputation so far has been insane. I thought it was just a mild curiosity, but it seems like it struck a nerve to know if it’s possible to live forever in a superintelligence (the element of comparison doesn’t hurt either!),” Dimson added.
While I’m not so sure that being “remembered” by a chatbot is a guaranteed ticket to immortality, I can’t deny that I find the results interesting and jealousy-inducing, especially since they’re codified in easily comparable scores. (AI critic Anthony Moser scoffed that this is “literally the equivalent of asking 13 chatbots to talk about you.”) It also helps that the site features a cute, Nintendo-esque retro design.
Dimson said he plans to further dig into why different models in the same series return different results, which models are biased toward different types of people, and which people “should have Wikipedia articles but don’t.”
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
