OpenAI’s Alexander Embiricos is coming to TechCrunch Disrupt 2026 — days after the launch of Dots

Alexander Embiricos, who leads product for Dots at OpenAI, is booked for TechCrunch Disrupt 2026 in San Francisco, October 13 through 15, only days after the company unveiled its always-on personal agent. His session is titled "What Comes After the Chatbot?"

On its own, a conference booking is a thin piece of news. What makes it worth a few minutes is the product it puts on stage. Dots get their own cloud computers, hook into the apps people already use, and keep working toward a goal in the background. That is a real step up in how much of a person's digital life an AI system can touch, and the person accountable for it is about to take questions in front of a room of founders and investors.

We are genuinely interested in agents that do useful work without being babysat. Handing off the dull, multi-step chores of digital life is one of the clearest wins AI could deliver for ordinary people, and a product that keeps going in the background is the logical next move after the chat box.

But the event framing itself names the hard part. How much control will people hand over? What should an agent be allowed to do without asking first? Can a company like OpenAI earn enough trust when the system learns more about us and gains access to more sensitive information? Those are not marketing questions. They are the whole ballgame for this category.

Here is what I'm watching for when Embiricos takes the stage. Does OpenAI describe concrete defaults, meaning what Dots will refuse to do without confirmation, what it logs, and how a person can see and undo what it did? Or does the conversation stay at the level of vision and vibes? An agent with its own computer and access to your accounts turns a bad answer into an action, not just a wrong sentence. If that scales across millions of users, the failure mode shifts from embarrassing chatbot output to money moved, messages sent and data shared that nobody quite approved.

We should also be honest about the setting. A conference stage is a friendly venue built to sell tickets, and nobody should mistake a fireside chat for an audit. The real tests of Dots will come from users, security researchers and, eventually, regulators. Still, putting the product lead in front of pointed questions this early is better than silence, and we'll judge the session by how specific the answers on permissions and privacy are.

OpenAI introduced Dots as an always-on personal AI agent shortly before this announcement. Disrupt 2026 runs at Moscone West in San Francisco, and Embiricos is slated for the event's AI Stage.