Anyone who has searched on Google recently knows that it doesn’t work like it used to. Of course, there are everything that happens with Google searchbut there’s also the inescapable feeling that web search is no longer the canonical source of information it used to be, and just as many people are learning who you and I might be through chatbots.
Thomas Dimson and Joey Flynn had a similar feeling, which led them to create on the weights. He “pesos” at issue are the numerical parameters that shape the training and production of an AI model, so the website intends to measure how well “a model is able to remember someone without using tools like web search.”
“Being in weights means that your existence was considered important in the process of creating superhuman artificial intelligence,” the website says.
To achieve this, In the Weights supposedly queries different models (including Grok, Gemini, multiple versions of GPT, Claude, and Llama, as well as lesser-known models) with a question similar to “Who is this?”
For example, this humble technology blogger I received a strength score of 641, putting me in the top 6% of names. I was feeling pretty good until I saw that. multiple TechCrunch colleagues got an even higher score. and the leaderboard has been changing as I write this post, with “Home Alone” star Macaulay Culkin currently in first place with a strength score of 988, neck and neck with opera singer Luciano Pavarotti.
The results also show which models returned which answers for a given name and highlight possible hallucinations; apparently GPT-5.4 Mini says Anthony Ha is an “ambiguous form of name that could refer to multiple people with the initials AHA.”
When asked why he created In the Weights, Dimson told TechCrunch via email that he and Flynn were looking to “get the creative juices flowing again” after leaving OpenAI (which they both joined through the acquisition of his design startup Global Illumination).
Dimson said he was thinking about how “personalized searches on Google are the wrong target in 2026 as more traffic goes to LLMs” and the fact that “so many lives are somehow encoded in a bunch of floating-point numbers inside the AI brain.” He also said that the address of the site was “sealed” by an ironic blog post riffs on AI weights and Terry Bisson’s classic tale “They are made of meat.”
“The reception has been crazy so far, we thought this would be a mild curiosity, but it seems to have struck a chord with us to want to see if you live forever in superintelligence (the comparison factor doesn’t hurt either!),” Dimson added.
While I’m not so convinced that being “remembered” by a chatbot is a guaranteed ticket to immortality, I can’t deny that I find the results intriguing and jealousy-inducing, especially since they’re encoded into an easy-to-compare score. (AI critic Anthony Moser he mocked that this is “literally the same as asking 13 chatbots to talk to you about yourself”). Also helping: the fact that the site features a cute image, Inspired by Nintendo retro design.
Dimson said he plans to delve into why different models in the same series give different results, which models are biased toward different types of people, and which people “should have a Wikipedia article but don’t.”
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.