The AI or generally intelligent AI that is like humans can be closer than we ever imagined. ChatGPT and its competitors can make very effective code very fast. Programming languages are things that are precise and easy to make especially for AI-controlled computers. The rules in programming languages are clear and precise. And that thing makes them easy to handle computers. The reason why programming languages are hard for humans is that we don't use them in any everyday communication.
The programming languages are full of symbols (like: <,$ or; etc.), that have no phonetic match in regular speech. Regular computers are handling data in binary mode, there are only two numbers, one and zero. One is on and zero is off. That is called a binary computer. Because there are only two numbers. That thing gives computers extremely high accuracy.
That thing is called precise logic. And that thing makes the data-based security keys effective. If binary codes don't match the key will not open the system. But a weakness of the precise logic is that every single action that the machine makes requires its module. If we want to give voice commands to the system we must use precise literal language.
If we use some dialects the computer doesn't understand what we say. The computer uses a "speech to text" application to transfer the speech to commands that the system drives to the command application. The fuzzy logic means that there is a level between the dumping application that can transform dialects into a literary language. It's possible to create a computer that understands dialects as well as literary languages.
These kinds of applications require only lots of databases, that they can transform any dialect into a literary language. The system can also transfer any speech, that used Finnish to any other language. The system transforms that speech into text and then turns it into a literary language. And translate it to English, and then it translates that speech to any other language like literary Japanese or its dialects.
General intelligence in the case of AI is not very hard to make. Every skill that the AI has requires its module. If we want to make an AI that can drive a car and translate Finnish to Japanese, we must just load the right modules into that AI. There are no people who can make everything. We must learn and practice things.
Computers are the tools that that learning otherwise. They just load the new operational module into it. Those translation skills and driving skills require their modules. And if the AI has access to the library where are those modules it can connect them to itself. The thing is that the binary systems are impressive tools.
But the next generation quantum processors are making the systems even more impressive than they are today. Maybe they use technology where computers are transporting data by using synthetic EEG. In that model, the system uses qubits to emulate the EEG. The thing that makes the human brain so powerful, is that they begin its operations in multiple points. The quantum system can use the internal qubit structure to make the human brain emulation.
The system can use the human EEG or brain electric curves to control its processes. The idea is that the system will just collect the EEG from humans who drive a car. Then the system must just compare the certain point of the EEG to data that it's connected from its environment. Then the system can just cut those EEG curves to a certain situation. And then the system can use that neural data to control vehicles and communicate with humans.
Comments
Post a Comment