Blog.

Essays, experiments, and practical notes on building with AI.

Cover Image for When "Sycophantic AI" Is Measured Through Reddit Morality

A study published in Science raises legitimate questions about chatbot advice. Its use of anonymous, reward-shaped moral scenarios raises separate questions about how much those results can actually prove.

Elliott A. Marquez
Elliott A. Marquez

More Stories

Cover Image for 7 Prompting Habits That Actually Help You Get Better AI Answers

7 Prompting Habits That Actually Help You Get Better AI Answers

Most prompt advice is too dramatic. Clear requests, better context, and a few simple habits usually matter more than prompt theater.

Elliott A. Marquez
Elliott A. Marquez
Cover Image for The Rise of Agentic AI Workflows

The Rise of Agentic AI Workflows

Agentic AI is spreading beyond chat and into real workflows, where systems can plan, use tools, and move work forward with increasing autonomy.

Elliott A. Marquez
Elliott A. Marquez
Cover Image for What is Vibe Coding?

What is Vibe Coding?

What began as a novel experiment is quickly gaining ground as a serious new standard in AI-assisted programming.

Elliott A. Marquez
Elliott A. Marquez