In the Weights launches AI-centric vanity search for personal brand monitoring
The AMW Read
Novelty is incremental as the concept of model recall probes is not new, but the specific consumer-facing product is novel. Significance is sub-segment as it primarily affects personal brand and niche interest, not enterprise-wide impact.
In the Weights launches AI-centric vanity search for personal brand monitoring
Thomas Dimson and Joey Flynn, former OpenAI employees who joined through the acquisition of their design startup Global Illumination, have launched In the Weights, a website that queries multiple AI models (including Grok, Gemini, GPT, Claude, Llama, and others) to determine how well a given name is "remembered" by these systems. The tool aggregates results, assigns a strength score, and highlights potential hallucinations, positioning itself as a modern alternative to Google vanity searches as web search loses primacy to LLMs for personal information discovery.
This product update exemplifies the recurring pattern of context-engineering moats and the shift from web-based to model-based information retrieval. In an ecosystem where foundation model labs compete on recall, accuracy, and hallucination rates, In the Weights exposes a new dimension: how individuals are represented in model training data. The service is a playful yet revealing probe into model biases, data curation, and the growing influence of AI on personal identity — a frontier that will increasingly matter for enterprise reputation management, hiring, and digital verification.
Dimson's observation that "Google vanity searches are the wrong objective in 2026 as more traffic moves to LLMs" captures a structural shift in information discovery. While the tool is currently a novelty, it foreshadows a market for AI-based reputation scoring and model recall auditing, especially as enterprises seek to understand how their brands, executives, or products appear in model outputs. The project also updates the open debate about whether model recall constitutes a form of digital immortality or simply reflects training data biases — a distinction with implications for privacy, copyright, and AI governance.