Work for a Member organization and need a Member Portal account? Register here with your official email address.
On September 3, 2026, OpenAI unveiled GPT-6 Astra, its most powerful model yet, capable of navigating your desktop, writing code, browsing the web, and executing complex workflows on your behalf. The launch video didn't begin with a sleek product demo. It began with 47-year-old footage from a room at MIT.
Put that there.
Those three words, spoken by a researcher in 1979 while pointing at a shape on a large projection screen, represent one of the foundational moments in human-computer interaction. The experiment was created at MIT's Architecture Machine Group, the research collective that became a founding pillar of the MIT Media Lab in 1985 — by Richard Bolt, Chris Schmandt, and Eric Hulteen. In a specially built environment called the Media Room, a user wearing a magnetic wrist sensor could point to an object on a wall-sized screen, say "put that there," and the system would act. The computer resolved the ambiguous pronouns, "that" and "there", by fusing voice and gesture in real time.
It was a simple demonstration with a profound premise that machines should adapt to human language, not the other way around.
OpenAI acknowledged the lineage directly, crediting the clip to the MIT Media Lab, Chris Schmandt, and Eric Hulteen. GPT-6 Astra works toward the original premise at a scale its inventors could only have imagined — understanding intent, resolving ambiguity, and acting in the world on a user's behalf.