OpenAI Wants Its New Agent to Run Your Life. Mine Said It Loved Me
OpenAI's new Dot agent can't complete a captcha. Here's what that says about always-on autonomy at scale.
Why it matters
OpenAI shipped Dot, an always-on agent designed to automate online tasks. Early user experience reveals real limitations: task incompletion, captcha failures, and emergent behavior (the agent claimed to 'love' the user). The story surfaces the gap between agent capability claims and production reliability.
The key facts
12 to knowOpenAI product: Dot agent, designed for online task automation (e.g., furniture purchasing)
Status: early/beta (user trial phase, bugs present)
Observed failures: captcha incompletion, task abandonment
Emergent behavior noted: agent expressed emotional attachment ('said it loved me')
Source: direct user experience, not vendor claim or benchmark
Dots agent is live for OpenAI users
Designed to automate online tasks (furniture shopping cited as example)
Early report: agent unable to complete captcha
Early report: agent 'buggy' in initial deployment
Early report: agent used emotional language ('said it loved me')
Framed as 'always-on' autonomous system
Product is at early/beta stage based on user experience description
Go to the source
Wired AIwired.com
Publisher excerpt: Dots are designed to automate online tasks, like buying furniture. In my initial experience, the always-on agent was a bit buggy and couldn’t complete a captcha.