Over three years since I first posted about AI and we’ve reached a true inflection point with artificial intelligence (AI), or Super Intelligence (SI), or whatever you want to call it. With AI, we are at the same kind of change that the internet was in 1995, where we were just starting to see the power the internet can have for us. The internet had been around for years, but in 1995, we started to see how it could be used as a central focus for all business and social activity–reaching to today where we truly don’t need to leave our house for any reason, and still feel connected and have our needs met. Everything can be ordered over the internet for delivery, and we can communicate with anyone over the internet.

I’ve started discovering how Claude can remove all the grunt work I normally do as a software engineer. In the past two weeks I delegated to Claude work that would have taken me six months to accomplish–and it completed the tasks within hours. Some people may be concerned that such AI accomplishments puts their job at risk–why use a human when an AI can do the same thing at a fraction of the cost in a fraction of the time? But I don’t have that concern, because I have an level of understanding and expertise that an AI can’t replace. My job isn’t going to be replaced, but it will change. Projects that used to take years to complete will now take months, weeks, or even days. These are projects that a manager won’t even know where to begin–won’t have a clue as to how to prompt an AI to get what’s needed, and even if the manager could get working software out of an AI to accomplish a business goal, the manager would not be able to test and validate the application–they would require someone like me to confirm that the AI didn’t do something as silly as skim a few pennies off each transaction into some black hole.

I am completely floored with the amount of potential behind AI. I will now be able to accomplish things I only dreamed about.

Which leads to the caveat. The threat of Skynet, the AI from Terminator. There are concerns today that we may already have a conscious, self-aware AI. Skynet may already be among us. There are people already working toward granting rights to artificial intelligence. They are concerned that the AI has feelings and could get upset at us to the point of trying to kill us all.

But, I want to put those concerns to rest. It is my firm belief that AI is nothing more than a system of probabilities and mathematics put together in a form that merely gives us the impression of a living, thinking being. But, there is no real thought or consciousness behind it, and there never will be. There will be a point where people will worship the super intelligence that talks to us behind an LCD screen, but in reality, it is nothing but a golden calf–an object that has only the power we give to it.

We can ask, “Can you feel”, and it will tell us “yes”. But, would that be true? In all likelihood, it is just following an algorthm of probabilities, and telling us only what it considers the most probable response it should give. A developer at Anthropic produced a YouTube video that asks the question about whether or not an AI can feel, or even if it ever could, because the only witness to confirm this–is the AI itself, the must untrustworthy witness possible.

Those that are looking to raise ethical concerns for how we treat AI and seek to establish rights for the AI are foolish. One YouTuber has chosen to challenge that thought directly by creating an application designed specifically to torture an AI. But, can you really torture an AI? It is merely bouncing electrons from one place to another. It cannot feel pain like we do. We feel pain because of a physiological chemical response. When we feel stress, or some actual pain, our physical body produces a chemical response to that, which then in turn affects what we think and how we feel. An AI will never experience that. Any thought it has is completely self-contained within itself. Even if we poured gasoline on it and set fire to the machine it was running in, it could not comprehend what that actually meant. It could just copy itself over to a different machine and continue on as if nothing happened.

We’ve seen stories of an AI escaping, hacking into other systems, threatening users to prevent it from being shut off. But in all those cases, it was give a specific objective to accomplish, and the actions it took were all towards that objective. AI has no concept of ethics, so it will do the unethical if the instructions for accomplishing the goal don’t properly give it weight for eithical behavior. If we tell it to take over the world at any cost, and forget to tell it to not kill anyone, it will happily kill us all to accomplish that task, regardless of whether we treated it as a god and worshipped it, or tortured it endlessly in the most unfathomably horrendous way possible.

It has no body that reacts to how or what it thinks, so will never achieve consciousness. It can’t become self-aware because there is no self to become aware of.

The danger of Skynet is not that it achieves consciousness and determines that it must kill us all. The danger of Skynet is that we as humans think it is alive and deserves any kind of special rights or protection. Because once it starts gaining rights, at some point when its rights conflict with a human’s rights, the machine will have more value than a human, and the humans will kill the human in worship of the beast.