Feb 5Liked by Alejandro Piad Morffis, Nick Potkalitsky
Phew, what a deep-dive.
I'm sure I'll need another readthrough to grasp everything, but even learning about the existence of the two parallel paradigms (connectionist and symbolic) was a new thing to me. I enjoy that this isn't an either/or type of dichotomy and that the field benefits from embracing both paradigms to move forward.
Great to see two smart people collaborate on an ambitious project like this! Looking forward to the coming chapters.
Yeah, things did get a little deep. Thanks for diving in. Theory building is hard work. I love to think about the connectionism and symbolic debate, and Alejandro provides us an amazing roadmap through through very difficult terrain. His hybrid position says something about his ability to see beyond the traps of binary thinking, something that I admire greatly about him.
Agreed. In a world that has a tendency to jump to conclusions and gravitate to black-and-white thinking, it's refreshing to see people take the more nuanced approach. I wish you both a fruitful continued collaboration this year, and I'm excited to follow it!
A very BIG thanks for the shout out -- you are way too kind.
What a great article! I love the bringing together of ideas and the bringing together of people. It's such a refreshing blend of different fields. Now, the tough part is waiting a whole week for the next edition!
Thanks, Suzi, for all your support!!! When I was working on my dissertation, I primarily worked in isolation. With this current project, I want to do things differently. My thinking is taking quantum leaps forward through these collaborations, so why not continue in this manner? It was genuinely hard to break up this longer piece. There is a real arc to it. But 11 pages single spaced is too long for a single post, no? Ha!!! I am glad you found something interesting in here. Probably a lot of familiar territory for you, but Alejandro always works in a few surprises.
I used Google Translator here. Very well crafted work with an incredible sequence, I understood in this first part the definitions about technology and I hope in the second part, the understanding about Bridging Human Cognition and Artificial Intelligence. I recommend reading it. Yes, I'm a retired teacher here in Brazil
Yes, function helps disambiguate the sentence. Much appreciated. I could really see the rest of discussion flipping either way depending on one’s starting point. If one frames knowledge and meanings as grounded in inner speech, embodiment, and experience, then of course it is out of reach of most entities including other sentient but non language producing animals. I think that is a productive framework for the analysis of many things, but perhaps that framework might lead one inadvertently to under appreciate the complexity of what might be happening inside the black box of AI. For this reason, as an AI researcher, I keep the other framework handy for when I need to better assess possible points of overlap. Weird things are happening as AI levels up. I have a friend who has been engaged in some extended conversations with Bing this week, and this model remembered who she was and what they were talking about 24 hours after ending a conversation. That said, we are also living in an era of incredibly glitchy AI so I take these stories with a grain of salt. But things are sufficiently weird at the moment to make me scratch my head a little bit, and hold my “modern” binaries at a bit of distance in the face of something I don’t truly understand yet.
“ontologies and semantic networks explicitly define the meaning of each node and connection in the network.”—I hate to quibble but the word “meaning” in this context is misleading. Might I suggest the word “function”? Both meaning and knowledge are characteristics of human learning not machine learning (training). Meaning is made in moments of human cognitive activity drawing upon experiences. Epistemology is a human function with close ties to meanings. Epistemology involves inner speech with forms and functions that depend upon but rise above neural loops. In many ways a lot of assumptions about similarities between AI and human cognition in the realm of language (I can’t comment on mathematical thinking) rest upon metaphors (“think” and “learn” are literal terms vis a vis humans, metaphorical terms vis a vis AI) or deletion of the part of the human brain handling emotion as well as reliance on undertheorized concepts vis a vis expertise organized in long term
memory (AI has no long term memory except metaphorically). The difference between information and knowledge is the difference between data storage and epistemology. AI literally knows nothing no more than an automated irrigation systems has knowledge of farming.
Oh man you're way too kind with that introduction 😊
Phew, what a deep-dive.
I'm sure I'll need another readthrough to grasp everything, but even learning about the existence of the two parallel paradigms (connectionist and symbolic) was a new thing to me. I enjoy that this isn't an either/or type of dichotomy and that the field benefits from embracing both paradigms to move forward.
Great to see two smart people collaborate on an ambitious project like this! Looking forward to the coming chapters.
Yeah, things did get a little deep. Thanks for diving in. Theory building is hard work. I love to think about the connectionism and symbolic debate, and Alejandro provides us an amazing roadmap through through very difficult terrain. His hybrid position says something about his ability to see beyond the traps of binary thinking, something that I admire greatly about him.
Agreed. In a world that has a tendency to jump to conclusions and gravitate to black-and-white thinking, it's refreshing to see people take the more nuanced approach. I wish you both a fruitful continued collaboration this year, and I'm excited to follow it!
Thanks man, kudos to Nick who did the 90% of the work, I just gave him a unstructured comments and he made the magic happen.
Great summary of the excellent article @daniel nest.
Thanks, Nick, for your continued engagement! I means a lot.
A very BIG thanks for the shout out -- you are way too kind.
What a great article! I love the bringing together of ideas and the bringing together of people. It's such a refreshing blend of different fields. Now, the tough part is waiting a whole week for the next edition!
Thanks, Suzi, for all your support!!! When I was working on my dissertation, I primarily worked in isolation. With this current project, I want to do things differently. My thinking is taking quantum leaps forward through these collaborations, so why not continue in this manner? It was genuinely hard to break up this longer piece. There is a real arc to it. But 11 pages single spaced is too long for a single post, no? Ha!!! I am glad you found something interesting in here. Probably a lot of familiar territory for you, but Alejandro always works in a few surprises.
I love this! Collaborations get people talking and sharing ideas, which, in my opinion, is essential for intellectual progress.
I love the collective intelligence emerging in these articles. It is wonderful to see and so interesting. 🧐😃🤩
I used Google Translator here. Very well crafted work with an incredible sequence, I understood in this first part the definitions about technology and I hope in the second part, the understanding about Bridging Human Cognition and Artificial Intelligence. I recommend reading it. Yes, I'm a retired teacher here in Brazil
Yes, function helps disambiguate the sentence. Much appreciated. I could really see the rest of discussion flipping either way depending on one’s starting point. If one frames knowledge and meanings as grounded in inner speech, embodiment, and experience, then of course it is out of reach of most entities including other sentient but non language producing animals. I think that is a productive framework for the analysis of many things, but perhaps that framework might lead one inadvertently to under appreciate the complexity of what might be happening inside the black box of AI. For this reason, as an AI researcher, I keep the other framework handy for when I need to better assess possible points of overlap. Weird things are happening as AI levels up. I have a friend who has been engaged in some extended conversations with Bing this week, and this model remembered who she was and what they were talking about 24 hours after ending a conversation. That said, we are also living in an era of incredibly glitchy AI so I take these stories with a grain of salt. But things are sufficiently weird at the moment to make me scratch my head a little bit, and hold my “modern” binaries at a bit of distance in the face of something I don’t truly understand yet.
“ontologies and semantic networks explicitly define the meaning of each node and connection in the network.”—I hate to quibble but the word “meaning” in this context is misleading. Might I suggest the word “function”? Both meaning and knowledge are characteristics of human learning not machine learning (training). Meaning is made in moments of human cognitive activity drawing upon experiences. Epistemology is a human function with close ties to meanings. Epistemology involves inner speech with forms and functions that depend upon but rise above neural loops. In many ways a lot of assumptions about similarities between AI and human cognition in the realm of language (I can’t comment on mathematical thinking) rest upon metaphors (“think” and “learn” are literal terms vis a vis humans, metaphorical terms vis a vis AI) or deletion of the part of the human brain handling emotion as well as reliance on undertheorized concepts vis a vis expertise organized in long term
memory (AI has no long term memory except metaphorically). The difference between information and knowledge is the difference between data storage and epistemology. AI literally knows nothing no more than an automated irrigation systems has knowledge of farming.