AI entrepreneur Matt Shumer was in his bedroom when he suddenly heard people talking in his apartment. For a moment, he thought someone might actually be inside.
There was someone talking—but they weren’t people.
Shumer, founder and CEO of OthersideAI, had asked OpenAI’s new GPT-6 Astra to create a survival world inside Unreal Engine and populate it with human-like characters. Each character was powered by Astra and given the objective of working with the others to survive.
Then something fascinating happened. The agents began talking to each other, discussing tools, helping other characters and coordinating what needed to be done inside their artificial world.
This wasn’t proof that AI had become conscious or spontaneously decided to create its own civilization. But the experiment demonstrated something important: AI is moving beyond answering questions and increasingly learning to act.
Welcome To The AGI Era
Just days ago, OpenAI President Greg Brockman used remarkable language to describe what is happening: “Welcome to the AGI era.”
Whether today’s technology technically qualifies as artificial general intelligence will be debated. But for the average person, there may be a more important development already underway. AI is becoming an agent.
Ask a chatbot, “How do I accomplish this?” and it gives you instructions. Tell an AI agent, “Accomplish this,” and increasingly it can use computers and software to pursue the objective itself.
Agents can conduct research, manipulate documents and spreadsheets, navigate websites, write software and carry out multistep workflows. We aren’t just building machines that answer us anymore. We are building machines that do things for us.
And now comes the next step.
What Happens When We Give AI A Body?
For all the extraordinary capabilities of AI, it has largely remained trapped behind a screen. It can write a report but can’t carry your groceries. It can analyze a warehouse but can’t unload the truck. It can explain how to fix your sink but can’t pick up the wrench.
Robotics could remove that barrier.
Combine an AI agent capable of seeing, reasoning, planning and learning with a robot capable of moving through the physical world and something fundamentally changes.
The robot becomes the hands. The AI becomes the brain.
Instead of programming a robot to perform precisely the same movement thousands of times, imagine giving it an objective: “Clean the kitchen.” “Unload this truck.” “Stock these shelves.” “Inspect this factory.” “Mow the lawn.” Or perhaps, “Help Dad get out of bed and make sure he takes his medication.”
The machine determines the steps necessary to accomplish the task and then physically carries them out. That is a very different kind of automation.
var images = [],
index = 0;
images[0] = ““;
images[1] = ““;
images[2] = ““;
images[3] = ““;
index = Math.floor(Math.random() * images.length);
document.write(images[index]);
//done
