Even the best AI cannot generate a thing. It just recycles old ones. When we think about paintings. And other things that AI. The truth is that it only connects details from sources that somebody else made. In real life, it connects information or details from multiple sources. The number of sources determines how effective and smooth results the AI can create.
The weakness is that the AI doesn't understand what it does. Productive AI like Chat GPT selects a certain number of sources by using the search engine's results. Then the AI just select one or two paragraphs from those homepages and connects them to a new entirety.
And here is the danger. The AI doesn't know what is reading in those paragraphs. The thing is that the AI recognizes the paragraphs. But because it doesn't know what is reading in those paragraphs. The AI can choose something that doesn't belong to any topics in the text. That means AI is far away from human intelligence.
So do we need scholars in the time of AI? The answer is that humans still in a vital role in machine learning. The AI can even collect thesis from homepages that human is prechecked. AI can play chess better than humans because it can calculate possible movements so far ahead. The AI can fly aircraft closer to the ground than humans because it doesn't afraid.
It just needs two or four radars that can fly in canyons. And of course, GPS and map where is the profile of the canyon. The radar altimeter can measure the aircraft's distance to the ground. And side radars tell the distance to walls.
The nose radar sees the possible obstacles. But that thing is not anything there the AI creates new things. An interesting thing about AI is that self-driving car requires more complicated code than drone needs.
The learning AI can learn multiple canyon profiles. It just records the aircraft's route. When we make self-driving cars, they need much more variables than robot aircraft needs for a successful mission.
Every variable is an object that the system must recognize. The AI-driven car simply stops the vehicle and says that the driver must use the manual system in city areas. The AI can make many complicated-looking things, but all those things that AI knows must be pre-program in its memory. Otherwise, AI is ineffective.
But when we think about the future and autonomously learning AI, we must realize one thing. The human brain requires information that it uses for connecting memory blacks.
The problem is this. We don't know the precise moment when the data collection begins.
The human brain involves 86-500 billion neurons. Calculating those neurons is difficult because of things like the number of axons. And the ability to interconnect neurons together has a great effect on the brain. The number of connections between neurons is the same way important. As the number of physical neurons.
https://shorttextsofoldscholars.blogspot.com/
Comments
Post a Comment