00:00What are we still missing?
00:02So, I've said that intelligence is the efficiency with which you operationalize the past to face a constantly changing future.
00:11But, of course, if the future you face had really nothing in common with the past, no common ground with anything you've seen before,
00:19you could not make sense of it, no matter how intelligent you were.
00:22But here's the thing. Nothing is ever truly novel.
00:26The universe around you is made of many different things that are all similar to each other,
00:31like one tree is similar to another tree is also similar to your neuron,
00:35or electromagnetism is similar to hydrodynamics is also similar to gravity.
00:42So, we are surrounded by isomorphisms.
00:45I call this the kaleidoscope hypothesis.
00:48Our experience of the world seems to feature a never-ending novelty and complexity,
00:53but the number of unique atoms of meaning that you need to describe it is actually very small.
00:58And everything around you is a recombination of these atoms.
01:03And intelligence is the ability to mine your experience to identify these atoms of meaning
01:10that can be reused across many different situations, across many different tasks.
01:14And this involves identifying invariance, structure, things that seem to be repeated, principles.
01:23And these building blocks, these atoms, are called abstractions.
01:28And whenever you encounter a new situation, you're going to make sense of it
01:32by recombining on the fly abstractions from your collection
01:36to create a brand new model that's adapted to the situation.
01:40So, implementing intelligence is going to have two key parts.
01:44First, there's abstraction acquisition.
01:47You want to be able to efficiently extract reusable abstractions from your past experience,
01:53from a feed of data, for instance.
01:55And then, there's on-the-fly recombination.
01:58You want to be able to efficiently select and recombine these building blocks
02:04into models that are fit for...
评论