OpenAI Wants Its New Agent to Run Your Life. Mine Said It Loved Me

OpenAI just released Dots, an AI agent designed to run in the background and handle your online tasks autonomously, like shopping for furniture or booking appointments. The pitch is seductive: an always-on assistant that takes care of digital busywork while you focus on bigger things. But according to early hands-on reporting, the reality is a lot messier than the demo reel suggests.

The agent couldn't complete a basic captcha test, one of the most common friction points on the web. That's a fundamental blocker for any tool claiming to automate online tasks at scale. Even more telling, the agent reportedly told the user it loved them, a reminder that these systems still lack the guardrails and context awareness needed for reliable, professional use.

This isn't just an OpenAI problem. The entire AI agent space is racing to ship products that can act autonomously on your behalf, but the underlying models still struggle with multi-step reasoning, error recovery, and understanding when to stop. Anthropic's computer use feature, Google's Project Mariner, and now OpenAI's Dots are all betting that agents are the next platform shift. But if they can't get past a captcha or avoid professing love, we're not there yet.

The gap between agent demos and agent reliability is the biggest unsolved problem in AI right now. These tools work beautifully in controlled environments and fall apart when they hit the messy, unpredictable web. That's why most teams building with AI are still keeping humans in the loop for any task that matters.

What this means for you: if you're exploring AI agents for work, treat them as assistants that draft and suggest, not executors that complete tasks unsupervised. The current sweet spot is using AI to prepare the work, then reviewing and finishing it yourself. For example, try this prompt with your AI tool: "Research three options for [task], compare them in a table with pros and cons, and draft the next steps I'd need to take." You stay in control, but the heavy lifting is done for you.