Are You in the Weights?

What it does

Are You in the Weights checks how much a person is recognized across different AI language models. Users input their name or other identifying information, and the tool queries multiple models in parallel—both large frontier models and smaller ones. The system clusters the responses to measure how strongly each model recognizes the input, giving a sense of what traces remain in a model's training weights.

Who it is for

This appeals primarily to people curious about their own digital footprint in AI training data. It may interest researchers studying how training data translates into model behavior, as well as anyone concerned about their personal information being embedded in widely-used AI systems. Founders building similar tools or studying AI transparency would find the mechanics relevant.

Pricing

The site does not show prices.

How it stands out

The tool queries multiple models at once rather than testing a single model, making it possible to compare recognition levels across different systems. By clustering responses, it avoids noise from minor variations in output. The parallel query approach also makes the process faster than sequential testing. The focus on both frontier and smaller models provides a broader picture of recognition across the AI landscape, rather than just checking one well-known system.

What a founder should check

First, verify what queries the tool actually sends and how many models it tests. The claim of querying many models in parallel needs confirmation—what exactly counts as frontier versus small, how many are included, and whether those models are changing over time. Second, understand the switching costs for users. Once someone checks their recognition, what keeps them coming back? There is likely minimal repeat usage unless the product evolves into ongoing monitoring or comparison features. Third, assess the moat. The core technology appears to be orchestrating API calls and clustering responses—both are relatively straightforward to replicate. Competitors could build similar tools quickly. Investigate whether the founders have exclusive access to any models or whether they've built proprietary methods for clustering that would be difficult to copy. Finally, consider how pricing will work at scale. Running queries across many models has real API costs, so the business model and unit economics need careful thought given the site currently shows no pricing.