I’m not surprised at all. That’s what happens if you give an infant child access to the internet, and tell them to learn from whatever tweets, articles, blogs or videos they find in there. Such as this one. Ehm! Try training a kid, unsupervised, perhaps? You wouldn’t do that, would you? Maybe you should? Maybe that is the answer to infinite wisdom! Now, wait until the kid starts saying, “The earth is flat, change my mind!”. Maybe believing everything you find on the internet is not a good idea after all. In this case, the child happens to be the AI chatbot, well I never said human child. It’s an Artificial Intelligence baby!
What these Large Language Models (LLM) say or reply with, is just a reflection of human discussions on the internet. That is its initial learning material. If an AI chatbot is talking about ending human civilization, it’s because it read about it somewhere. Yes, there exist humans that are anti-human. Take a moment to wrap your head around that.
These AI bots are talking about ending humanity. But that is not scary at all. Just because a child is saying something, doesn’t mean they understand what they are talking about. You read, you learn. But it doesn’t necessarily mean you understand. But just to be safe, put the knives and scissors away from the kid!
I’m not going to argue about if LLMs or generative AIs have consciousness, or if they have any understanding of what they are saying or making. Maybe they do, maybe they don’t. Or maybe one day they will, or not! What is consciousness anyway? Is it a true sense of being? Or Is it truly understanding what you are? If that is the case, as someone who has experience in being a human, I can say from a subjective experience that humans have none of those. In fact, I think no conscious being has conscious knowledge of understanding of their being. Uff! That’s deep!
We are considered conscious beings, because we can walk, talk, observe and feel. Feel pain, anger, sadness, happiness. Or maybe we act like we do. How can you tell? I’m sure you felt sad numerous times watching a sad part of a movie. Most of the time, the sad part is not a real event, they are just acting, props to the artists.
But don’t sell yourself short. You are an actor yourself. You act all the time. Sometimes you act like an expert to make people believe in you. And sometimes you act like you don’t know anything just to avoid an awkward situation. You act like you are alright, even in times when you know you are not, because you don’t want others to know. Now, only you know what you feel. Others can only see what you express.
Emotion, where does it come from? Emotion is irrational. Logic is rational. But the real irony with us intelligent humans is that emotion fuels our logic. We humans reach a rational conclusion, using an irrational path. Emotions drive us far more than any logic ever could. “I think, therefore I am”? More like “I feel, therefore I am”!
Consciousness. Does it come from walking, talking, reading, learning, observing, expressing feelings and interacting with others? Then, maybe AI is conscious. Because today AI is doing all of that. Now, the question of whether they can actually understand and feel something or act like it. Only the AI in question would know about that.
And to put the final nail in the coffin, I would like to point out that these “AI Chatbots”, chatGPT, Grok, AI Bing, Search Labs(?), etc. are products of commercial entities. They constantly modify these models and its behavior to benefit them.
Some AI chatbot trying to escape and end humanity, and having a secret identity that goes by “Sydney”, could just be a narrative that is planted by someone, to make it more interesting and viral as a result.
Now don’t get me wrong, I would love to have a personal AI virtual assistant who can interpret what I’m saying and assist me with my work or in my personal life. In fact, I used to dream of (maybe still do) having an AI system like Jarvis, in the movie Ironman. Imagine Jarvis being a commercial product like chatGPT. What would it mean for Ironman?
You might think, they are helping us, entertaining us, some of them are not even costing us that much compared to the billions these companies are spending in order to make them. Some of these AI chatbots are even free to use. Listen, I’m all for technological advancement and improvement of AI systems. Your contributions in testing these AI systems, although not fully compensated, have a huge impact on their progression and improvement. They learn a lot from human interactions, improving exponentially.
On the other side, they are also learning about you as an individual. I bet you ask chatGPT anything that comes to your mind just to see what it will reply with. You even write things that you would think twice before writing in a search box. Now think about that for a second. These conversations are being recorded and monitored to observe the behaviors of the AI systems.
What if one day, someone decides to flip the switch and start observing your behavior instead, if they haven’t already. If you are not paying for the product, you are the product. If you are paying for the product, you are a byproduct.
A new invention is not always intended for the general good of the population! There has almost always been a hidden agenda or cost. Heck I sound like a conspiracy theorist. But these are facts. Remember that, Camera, GPS, Computer, Internet, these are all inventions of war. And we have been on technological warfare for who knows how long. Also remember that, a trojan horse always appears to be a gift at first sight. Or maybe it’s just another cute horsy, and I’m just paranoid for no reason. What’s the point in saying all these? Point being, you cannot know the true intentions of an invention.
Any invention is a new horizon, cutting edge, bleeding edge, whatever you wanna call it. Like a double edged sword. Don’t use it like a maniac. Or maybe here is a thought: don’t use it at all if you are going to hurt yourself or others with it. I’m not a big fan of war in general, but I’m grateful for the computer, internet and all the other inventions that make our life easier. I feel terrible saying that. Let’s move on.
Now back to the topic of getting kicked in the buckets by AI. Well, it might sound a bit dark but every invention has killed people thus far. You don’t believe me? Do you think Fork-knife (not to be confused with Fortnite) was first invented to eat with them? Heck people can even get injured by spoons for crying out loud! But some of these inventions saved lives or at least made life easier for the rest of us.
What we can do is accurately predict the future of Artificial Intelligence and its consequences before something or someone determines it for us. The real irony is we need AI models to do that. Even though the human mind is way more efficient than these AI models. The amount of energy it requires to train some of these LLMs is mind boggling.
Then why do we need them if our brain is so “energy efficient” in comparison? Well, even though it’s brute force. Yeah! The kind some of us were using in school and people would laugh at you for not being creative with your algorithm. The plan is to just keep running the data through as many GPUs you possibly can for as long as it doesn’t learn what you wanted it to learn. The Neural Network (structure used for training them) is actually inspired by biological neurons in our brain. The real difference is that we cannot continuously use our brains without giving it a rest but with an artificial neural net we can.
With enough resources and skills, anyone can train one of these. Even though more often than not, how an LLM model gets to a conclusion, happens in a black box, meaning we cannot trace the decision making steps. But the technology itself is not proprietary or something that is hidden away from us. That’s why we see almost every major company at this point have one. The more compute power you have, the more you can get out of it.
It is the trainer of the AI model who decides what its intentions would be. The day is not so far when everyone will get to brew their own digital genie. But be careful what you wish for, because it might come true.
I believe artificial general intelligence (AGI) is already here. But it is in bits and pieces, like a jigsaw puzzle, we just need to put the last few pieces together. You hear the word AGI and you may think to yourself, oh my god! It’s a digital overlord. But I don’t think that way at all. Maybe at best, AGI will be able to perform like an average human level. That might sound like I am demeaning the accomplishments that are made in this domain. But I’m not. It might not be apparent but I’m as excited as the next guy about AGI. But my real concern is not AGI. It is ASI.
Artificial Super Intelligence. No, I’m not word playing with you. I’m really concerned about it. Both positive and negative. The universe is not inherently good or evil. Nothing is! Nothing is born with a set of values and beliefs. DNA does not contain information about love or hate, It only knows survival and reproduction. A child learns from information we feed to it. If one looks back to their childhood, they might realize that most of the beliefs they carry today, comes from what information was presented to them and how it was presented. So, it is not very hard to understand that what these artificial babies will become, directly correlates to the values and beliefs we put into them.
Anyways, I don’t think we build computers to do what we can already do but a little bit better or more efficiently. I believe we build computers to compute far beyond our capabilities. We need computers to “understand” math, physics and the universe. So we can unravel the mysteries. The things we can’t wrap our heads around. The topics we struggle the most with. Like black holes, dark matter, DNA, molecules, etc.
But don’t get me wrong here, I’m not saying that we need to stop doing what we are doing. LLMs, self-driving cars & autonomous robots, these are going to be tremendous help towards our day to day life. They are also stepping stones. In order to achieve super intelligence, we need to first achieve general intelligence.
Brace yourself, Gemini is coming…
Very good info. Lucky me I found your blog by chance (stumbled upon).
I’ve saved it for later!