344
Productivity & Workflow355
Automation & Workflow224
Software Development250
Marketing & Growth192
AI Infrastructure & MLOps174
Writing & Content Creation203
Data & Analytics141
Design & Creative169
Customer Support131
Photography & Imaging156
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
A new website, In the Weights, checks multiple chatbots to estimate whether they can identify a person by name and gives a public score.
In short: A new website called In the Weights gives you a score meant to show how well different chatbots can identify you by name.
In the Weights is a new site created by Thomas Dimson and Joey Flynn, who previously worked at OpenAI. The idea is similar to searching your name on Google, but it focuses on chatbots instead of web search.
The site says it measures whether an AI model can “recall someone” without looking things up on the web. It does this by asking several major AI models, including Grok, Gemini, multiple versions of GPT, Claude, and Llama, “Who is <name>?” It then groups similar answers together and gives the name a “strength score.”
The site also shows which models answered, and it flags possible “hallucinations,” which is when a chatbot confidently makes something up. For example, TechCrunch writer Anthony Ha’s page shows a strength score of 641, and it highlights a strange response from one GPT model.
Dimson told TechCrunch the project came from a feeling that “Google vanity searches” matter less in 2026 as more people learn about others through large language models (chatbots that predict text, like an autocomplete on steroids). He also said the reception has been strong, partly because the score is easy to compare.
As chatbots become a common way people get information, they can shape reputations the way search results once did. Tools like this can make it easier to spot when chatbots are wrong, but they can also encourage people to chase a number that may not reflect real-world importance.
Source: TechCrunch AI