AI = Making Distinctions
By Leo Gura - April 5, 2023
I have an episode called Learning = Making Distinctions, in which I explain how to super-charge the rate and depth of your learning by consciously focusing on making sharper distinctions.
If you study the mechanics of how these powerful new AIs are trained, it’s exactly by this method! AIs are trained using sharp, clear, distinct data-sets. For example, you feed an AI 1000 clear images of dogs and 1000 clear images of cats, and it will learn how to distinguish them. If the data-set is murky or ambiguous, the AI will not learn well. Your mind works in the same way.
One of the things that makes learning anything difficult is that no one gives you enough clear, sharp, unambiguous examples. Much of the labor of learning is finding these distinct examples out of all the crap humans feed you. One of the biggest parts of AI research and development is to use humans to curate very clear data-sets. For example, if you want to train a Tesla car to avoid driving over children, you have to curate a data-set of 1000s of videos of children running across the street. Then it will learn to recognize or distinguish children. Otherwise it won’t.
Which is why I fill my episodes with dozens and hundreds of excellent examples. I rack my mind to come up with the best examples for you. Because I know I need to train you like a neural network.
A neural network is a massive set of distinctions that mirrors the infinite set of distinctions that we call Reality.
I just explained the metaphysics of AI.
Click Here to see ALL of Leo's juicy insights.