OpenAI turns its coding tools into dots, an always-on personal assistant

OpenAI has extended the capabilities of its coding tools into a product called dots, an always-on personal assistant, according to TechCrunch.

The report is built around Alexander Embiricos, the OpenAI executive responsible for dots, whose bluntest line is that he never wants to have a meeting with someone else's AI agent.

That line points at where the product is heading. An assistant that is always present does not just answer individual prompts; it sits between the user and other systems, including other agents acting on someone else's behalf.

Starting from coding tools makes a certain sense. Code is the domain where agentic behavior is easiest to check — tasks have clear boundaries, outputs can be tested, and failures can be rolled back — which makes it a practical place to harden a system before pointing it at open-ended work.

Positioning dots as an always-on assistant is a different bet from a chat window. It moves the model from something a user invokes to a layer that stays in the background, which changes the questions about attention, permissions and trust.

The open questions are the product's boundaries: how dots is distinguished from OpenAI's existing assistant, how it stays present without becoming intrusive, and how authority is granted when it acts for a user in front of another company's agent. Embiricos's remark about meetings suggests cross-agent interaction is exactly the part being designed for — and the hardest part to standardize.