Summary: Build AI products around evidence, not ideas: prototype quickly, test with agents and real users, and iterate until the product proves its value.
Treat AI advice, models and agent workflows as hypotheses: test them on your own work, mine your logs for evidence, and make verification part of the workflow.
I found Idea-a-Day, a site that releases a copyright-free concept every day. It's a great resource if you need a quick spark of creativity or an open-source project idea to build upon.
I challenged several LLMs to generate funny Tamil puns ending in ".ai". While DeepSeek and Claude struggled with accuracy, Gemini captured the cultural nuance perfectly with clever wordplay like Tholl.ai (annoyance) and Kaval.ai (worry).
I share how I've used AI in my Tools in Data Science course: teaching through exams and prompts, designing assessment to make cheating pointless, building course content, and analyzing learning data.
