Greg had Gemini 4 Argon rebuild his holding company's site, write a post in his voice, and build a Shopify product customizer that exports 3D-printable STL files.
Predict customer churn from a database table without training
Prompt Engineering builds a TabFM-Auto agent on Neon Postgres that predicts customer churn zero-shot, and shows a leaky column fooling it and a database branch fixing it. Sponsored.
Prompt Engineering · YouTube · published Oct 11, 2026 · Research & Knowledge
▶ Watch on YouTubePrompt
Using the rows in [customers table] where churn is known, predict churn for the remaining customers. Check first for columns that leak the answer and exclude them.
How to try it
- Open TabFM.
- Copy the prompt and fill in the [placeholders].
- Watch Prompt Engineering's video for the full walkthrough and setup.
Related use cases
Cole built an agent that knows his interests, reads hundreds of sources daily and sends one short morning digest, then shares it as a free skill that interviews you first.
The free Last 30 Days skill scans YouTube transcripts, Reddit, Polymarket, Hacker News and GitHub, then scores and condenses it into a short report instead of stale web search.
Leo has the n8n assistant, powered by Gemini 4 Argon, build a quote request workflow that scores leads, drafts email replies, stores results in a data table and pings Slack for urgent ones.
Share
Show a real task done with your tool in your own video; we write it up and label it Sponsored.
Example · Sponsored$149 one-time