Project lead and author
Aaron Decker
Aaron leads Hallucination Yield, an independent research project studying the investment opinions generated by AI models and the real-world infrastructure and supply chains behind the AI economy.
Research approach
Hallucination Yield began with a simple question: do large language models systematically favor particular public companies or overestimate their importance? The project records model responses over time so those preferences can be measured instead of treated as anecdotes.
The research now also follows company-reported infrastructure spending, official Taiwan monthly revenue, semiconductor price and trade series, and public portfolio disclosures. Each dataset keeps its source definition, unit, period, provenance, and important limitations.
Research is published for education and investigation. It is not financial advice, and neither Aaron nor Hallucination Yield warrants that a source, model response, interpretation, or derived metric is complete or accurate.