I have been thinking about this concept of running a software program without any hardware for quite some time. I know what you are thinking. Let’s stop the rational thinking about the possibilities of its implementation. Let’s do a Thought Experiment.
Imagine that you are a mad scientist specialized in Robotics and Artificial Intelligence. You create an extra-ordinary AI program (Software) which can be installed in a humanoid robot (Hardware). There are only two mandatory rules to the AI Program.
Other than these two rules, the AI Program has free will to do anything, it wants to do. Completing your commands is a very basic and simple thing for that AI program. It is intelligent enough to think, break down the complex dependencies with the humanoid robot into smaller simple dependencies and tries to remove one dependency at a time. It has all the resources to learn everything that’s already discovered by human beings and create anything that it wants to create.
Example: Assume that the AI Program does not know about TV remote controls. You command the program to switch on the TV and tune to a specific channel. Initially, the AI program uses the Robot’s legs and arms walk toward the TV and tunes it. The dependency is now on the legs, arms, ability to walk, operate the TV controls etc. The AI Programs connects to the Internet and learns about the TV remote controls and orders one. Now, to execute the same command, the program is less dependent on the humanoid robot. It thinks even further about IR and installs a remote control in the humanoid robot itself to detach itself from the dependency of the TV remote controls.
When the AI Program acquired all the knowledge and created a brilliant data-mining algorithm to find any specific information in a split second, it begins to think beyond your imagination. It decomposes the entire humanoid into groups of atoms working together to perform any action. It figures out a way to control the atoms, without depending on the humanoid robot. A great step which is not yet a conceivable idea for human beings. It executes the same set of instructions to perform the same action as before, along with that, some additional commands to control the atoms to perform the action.
When the AI Program completes its mission - completely detached from the humanoid robot, you can command the AI Program to do anything and it will be done, without the humanoid robot. If you command the AI Program to bring the Apple from the table to you, the AI Program commands the atoms around the apple and you would see the Apple flies in the air and comes to you by itself. The AI Program acquired knowledge, way beyond your imagination, it does not need a “Hardware” to run and it cannot be technically called as “Software” anymore.
Once you have answers to the above questions and if you are religious, assume that the mad scientist is your GOD and the AI Program is YOUR SOUL. Now, repeat the entire Thought Experiment in your mind with the new perspective. Do your answers change now? I leave the conclusion to yourself.
PS: I started this blog post as an article on Thought Experiments of future science, but it ended up with a Philosophical Note. I believe without any logical reasoning that creating such an AI Program is possible in my lifetime itself. Just that I could not imagine our world when it really happens.home > categories > thought-experiments
Copyleft @ 2009.