Former OpenAI employees Thomas Dimson and Joey Flynn have launched In the Weights, a free website assessing how well AI models recall information about individuals. Dimson predicted Google searches for oneself would lose significance by 2026 as information flow shifts to language models, according to Zamin Uz. People currently rely on traditional search engines to understand their online presence, but In the Weights reveals an individual's digital footprint is increasingly defined by the internal parameters of AI models. This creates an opaque form of digital identity beyond traditional search. Therefore, individuals and companies will increasingly need to monitor and potentially influence their representation within AI training data, fundamentally shifting the battleground for reputation management and personal discoverability.
How 'In the Weights' Works
- The In the Weights tool measures whether an AI model can identify, describe, and contextualize a person without performing an external search, according to Startup Fortune.
- It assesses AI models such as GPT-5.5, Claude Opus 4.8, Gemini, Grok, and Llama.
- The service queries these models with the question 'Who is this person?' to analyze responses and provide a 'strength score', according to Zamin Uz.
By directly querying the internal knowledge of leading AI models, In the Weights offers a novel, direct insight into an individual's 'AI footprint.' This method bypasses conventional web indexing, revealing a person's digital presence as defined by internal model parameters rather than external search results. The tool's 'strength score' quantifies this internal recall, providing a new metric for digital identity.
The Shifting Landscape of Digital Identity
Thomas Dimson's prediction that Google searches for oneself would lose significance by 2026 signals a profound reorientation of how individuals discover and validate personal information. This shift moves away from public web search results, instead emphasizing AI's internally defined knowledge. This fundamentally alters who 'owns' one's online persona, transferring influence from indexed web content to the opaque training data of large language models. The implication is clear: managing one's digital reputation will increasingly involve understanding and potentially influencing these internal AI representations.
The emergence of tools like In the Weights suggests that the future of personal discoverability and reputation management will likely hinge on an individual's presence within AI training data, rather than their traditional search engine ranking.
