Key Takeaways

  • OpenAI Dots run as persistent, always-on agents operating across ChatGPT, Codeex, cloud instances, local machines, Slack, and phone calls.
  • Product executive Claire Vo tested a custom Dot named Bae, an agent with its own dedicated computer built for asynchronous tasks like coding and scheduling.
  • The product suffers from severe interface friction because users cannot easily distinguish between a standard chat thread, a Codeex thread, and an active Dot thread.
  • For actual daily work, Vo continues using specialized micro-agents inside Grok rather than relying on OpenAI's broad autonomous agent.

The Multi-Channel Reality of Autonomous Dots

At OpenAI DevDay 2026 in San Francisco, OpenAI introduced Dots to push past single-prompt chat windows. The goal is an autonomous, persistent worker that lives across every platform you touch.

Vo tested the architecture directly. As she put it, “Dots are an always on longunning agent that can work across chat GPT across codecs in its own computer on your computer on your phone whether through text or through call even in your Slack to get work done for you.”

To see how this works in practice, Vo created her own custom agent. “This is my dot. Her name is Ba. She's a pink blob with glasses and googly eyes. And Bae has her own computer.” Bae can pick up requests, execute multi-step coding jobs, manage calendar items, and make external phone calls or Slack updates while the user is offline.

Giving an AI model its own cloud instance changes the operational dynamic. Instead of waiting for a streaming response, you hand off an objective. The agent logs into its dedicated environment, tests code, checks sites, and messages you across whatever channel you are currently checking.

Where OpenAI Thread Architecture Breaks Down

While the underlying execution works well on coding and purchasing tasks, the user experience introduces friction. OpenAI has layered Dots on top of existing ChatGPT and Codeex products without resolving how these primitives interact.

Vo highlighted this core product tension: “I think there's a bunch of rough edges around dots position in all the primitives that you have going on with chat and codeex. I think it's very confusing to understand what's a dot thread, what's a chat thread, what's a codeex thread.”

When you open your workspace, you are forced to make a cognitive choice before typing a single prompt. Is this quick question meant for a lightweight chat? Does it require the developer tools inside Codeex? Or should it spin up a persistent Dot that runs in the background? If a Dot executes a script, does that context stay in the Dot's history or sync to your standard chat sidebar? OpenAI does not yet provide clear answers to these questions.

Because of this ambiguity, Vo is keeping her existing workflow intact. “I will say right now early take and again I'm not doing a full dot review. I'm still sitting and using my Grock bots for work. I just like defining microaggents for work just like I would hire someone.”

What to Do With This

Do not rebuild your team workflows around a single generalist agent right now. Map your top three repetitive weekly tasks, then write narrow prompt definitions for each one as standalone micro-agents in your current tool of choice. Treat each agent like a specific junior contractor hired for one exact job, rather than expecting one omni-channel assistant to manage your entire software stack.