I Let Dots, Muse, and Spark Take Over My Daily Chores. Only One AI Passed the Test

I’m not a big fan of AI agents. They rarely complete the tasks I assign them, and even when they do, they take longer than I would. But with a new crop of agents making headlines (and adopting deceptively cutesy mascots), I decided to revisit my conclusions and test ChatGPT’s Dots, Gemini’s Spark, and Meta’s Muse across a series of everyday tasks, including adding groceries to an online cart and reducing email spam.