Introducing dots
OpenAI is rolling out “dots,” persistent agents with separate cloud computers, connected apps, background work, and approval controls for consequential actions.
The launch describes read-only proactive research, configurable permissions, activity tracking, and an auto-review step for actions that could affect accounts or share information. OpenAI says sensitive tasks such as changing passwords remain with the user; it also warns that dots can make mistakes. These are product and safety claims from the vendor, not an independent evaluation of the system’s safeguards.
HN commenters questioned how dots differ from Codex and ChatGPT Work, and debated whether an always-on assistant merits a separate product and how it may be priced. The discussion also raised the tension between consumer convenience and stronger workplace controls.