00:00:00 Introduction 00:05:02 Cognition space 00:07:52 Are there general rules for complex systems? 00:09:40 Evolution of cognition 00:12:58 Single-cell cognition 00:15:10 Where do cells store information? 00:17:10 When the first neuron came to be? 00:20:41 Inter-neurons allow Information Processing 00:24:27 Von Neumann Architectures vs Neural Networks 00:27:55 Evolution of Brains and Computers 00:30:18 How AI is different from human cognition? 00:33:25 The brain at the edge of criticality 00:36:22 Phase transitions in the brain 00:38:39 Liquid brains: brains formed by individuals that move around 00:42:40 Solid vs liquid brains 00:44:35 Is an ant colony a brain? 00:46:34 Slime mold solving mazes 00:49:52 Plants as brains 00:51:45 Brains as multi-layers of threshold elements 00:56:27 Constructing artificial cognition 00:58:09 Symbolic vs connective approaches of AI 01:00:00 Embodiment of AI 01:02:55 What can we build in the cognition space? 01:05:25 Filling voids in the cognition space 01:07:04 At what point society becomes a collective intelligence?
Im really curious about this topic despite not knowing a lot about it in a deeper sense. I have a gut feeling there's a lot more to this than we currently understand and appreciate. A topic to keep an eye on...
Some basic questions: Are there "thresholds" in complex systems, where emergent properties appear? Do they have something common in general? Does physicalist Strong Emergence exist? ( in the mild sense: not new physics that affects the basic, fundamental laws, but S.E. in the sense that some effects and emerging properties cannot be deduced, even in principle, from the fundamental laws)
That was quite interesting. If I heard correctly, I think he said that the Senses are "interviewers." I would have to disagree with that, as they seem to be unbiased "Reporters" that received data, and convert that into information that can be communication within the limitations of their capabilities and system. Near-sighted eyes are not good reporters of distant objects, reporting out of focus and distorted images which reflect both the limitations of that person's vision, along with the degraded data.
1:00:55 is scary part is giving them weapons, which police departments want to do with Boston dynamics. AI should never be armed or controlling major infrastructure. They should advise, alert, summarize, teach, but never be put in control.
I contend that a Tesla car with FSD (Full Self Driving) has a theory of mind, allowing it to predict the potential actions of other road users and act accordingly. This could show the benefit of an embodied AI.
i have no familiarity with self-driving cars particularly, but could we suppose it a theory of car? id imagine you'd be learning car states in relation to whatever else which would give you a theory car, which could scale to a theory of mind.
@@connorkapooh2002 I posit that Tesla FSD (Full Self Driving) has been trained to anticipate the actions of other road users. There are examples of the Tesla hesitating on a narrow road for a car coming the other way in case it pulls out. It also anticipates the actions of pedestrians. The Teslas behaviour is innate. Tesla trained FSD version 12 entirely from carefully curated examples of good human driving. It is being released to the entire fleet for a month.
@haroldmatias12 It does more than that. It anticipates what other road users could do and acts accordingly. If nothing else, it needs to distinguish between animate and inanimate agents. Being able to anticipate potential actions by pedestrians and other drivers demonstrates a theory of mind in my view. FSD demonstrates this ability by driving without interventions through suburban and city traffic.
