Investing Unscripted 188. How to Use AI to Research Stocks (w/ Brian Feroldi)
9 snips
Jan 14, 2026 
Brian Feroldi, founder of Long-Term Mindset and a seasoned investing writer, shares his insights on enhancing stock research using AI tools like Gemini. He explores the difference between price and earnings bubbles, using Nvidia as a case study. Brian discusses effective prompt strategies to minimize AI hallucinations and even runs a live demonstration analyzing Green Brick Partners. He emphasizes the importance of clear instructions when working with AI, likening it to an eager intern ready to learn.
AI Snips
Chapters
Transcript
Episode notes
Two Types Of Market Bubbles
- There are two bubble types: price bubbles and earnings bubbles, and earnings bubbles are harder to spot.
- Brian worries AI-driven capex could create an earnings bubble even if standard valuations look reasonable.
Insist On Source Links
- Always ask the AI to provide links to original sources like SEC filings or transcripts.
- Verifying those links quickly boosts confidence that AI outputs match the primary documents.
Role Assignments And Stepwise Prompts
- Assign the AI a role (e.g., value investor, forensic accountant) before analysis to shape its perspective.
- Give step-by-step instructions so the model knows exactly which documents and metrics to extract.