You’re trying to solve the wrong problem.

I was watching a video the other day of a robotic filling station. They programmed a robot to open a gas tank, then take a standard gas nozzle and insert it into the tank, select the gas, and then finally fill the tank by operating the handle. While it was an interesting video, they solved the wrong problem. The problem was how to put fuel into a car. From there, they assumed they couldn’t change the pump itself. While it is a reasonable assumption that you could not change the car, why did they assume that they couldn’t change or modify the pump?

Too often, when we are trying to use robotics or AI, we are trying to make the robot or AI fit the existing ‘framework’. When we want to use robotics or AI to solve the problem, we must step back and understand the problem and, more importantly, our assumptions.

It takes a lot to see our unconscious assumptions.

A friend of mine said that we are trying to change the alien intelligence to fit the framework. The framework was built for humans now, we are trying to make a robot like a human. While the robot and AI are changeable, so is the infrastructure/framework. The robot and AI can be integrated into the framework. One solution to putting fuel in the car would be to look at doing something similar to a tanker aircraft. Some of them use a maneuverable boom that can be controlled.

The question is, ‘Can we change the pump’?