What do people really think about AI?
Fast directional reads from a synthetic audience grounded in US population structure. Not real surveys — but honest about the uncertainty, and we never publish a result without genuine spread.
Do cold sales emails actually work?
Everyone gets cold emails — but has anyone ever acted on one? A directional synthetic read of US working adults.
77% said “No, never”
Read the study →Would people trust a decision made from AI-simulated opinions?
Synthetic respondents can pre-test an idea in minutes — but would people trust a decision made that way? A directional synthetic read.
78% said “Only as an early signal, then verify with real people”
Read the study →Do people trust AI answers more than Google?
As people ask ChatGPT instead of Google, who do they actually trust more? A directional synthetic read of US adults.
66% said “No — I trust Google results more”
Read the study →Method: a language model is conditioned on real US age and sex structure (World Bank) and answers in character. Results are directional, not real surveys. Studies where the model collapses to a single answer are never published.