Replies

I suspect that LLMs based on deep learning do capture much of consciousness but critically not all as Matt notes.

It certainly could be said that humans themselves are nothing but predictive machines. For what it’s worth I think the big tech/prosecutor/finance/real estate/military industrial complex seditious conspiracy organized by the Judicial branch and their partisan dualist political parties would have it be so. This difference I think highlights Matt’s point.

I would like to understand more about the deep learning aspects of large language models. A process educated specialist could likely right this in an accessible fashion. I am however aware of no accessible ‘deep’ source from any perspective.

I have read Sejnowski’s ‘Deep Learning Revolution’ and found it worthwhile.

https://en.wikipedia.org/wiki/Terry_Sejnowski

For what it’s worth I’ve started using Google’s MetaAI, via the Messenger interface. I suspect it’s not quite as good as GPT4, but it is free. It will also do simple image generation with the instruction ‘imagine’. It is very easy to use.

As a timely aside last month’s long running Gallup Poll on political affiliation had independents for the first as an absolute majority, bigger than both parties combined at 51%. In recent history the number has been in the 40’s, and frequently a plurality. I suspect that trend will continue.

I forget where, but I recall Matt making reference to analytical philosophy turning toward panpsychism. It’s an important question, but I have only so much time, and mental energy, to prehend.

Does this interview go to this?

That is a nice summary. One thing that I’ve found in life is the enrichment arising from a diversity of perspectives. That certainly applies to process thought.

We all bring to the table unique minds that see these questions at least a little bit different, and by coming together we all see more.

I am reminded of the parable of the blind feeling different parts of an elephant.

‘The Rythym of Becoming’, that title says so much just on its own. In it I find a possible, at least partial, answer as to how experience accumulates at the molecular level.

I’ll chime in with appreciation as well. For what it’s worth, I suspect the power of good poetry stems from its musicality, and all that this means.

A tangential question from the first discussion session organized by Chris Hughes: Is wonder a feeling, an emotion, in the whiteheadian sense?

Simplifying, how does experience accumulate at the molecular level?

I’ve just recently read Emerson’s collected essays, the essay on nature is very whiteheadian.

For what it is worth I do not understand why this book is not required reading in high school, it is so fundamentally relevant to the American experience.

This is an important topic to me, but my deep focus is elsewhere for this course. I’ll put that book on my list. The subject would make for a great Cobb course.

Actually the deep learning model of AI, claimed to be the most successful, is an attempt at modeling neural networks. I am not an expert, and I don’t know specifically on poetry functions.

This article confirms, and also raises some questions: https://news.mit.edu/2022/neural-networks-brain-function-1102

My source is Terrence Sejnowski whom I’ve heard speak and his 2018 book: https://mitpress.mit.edu/9780262038034/the-deep-learning-revolution/

Would you have an accessible source for more info on category theory?

Sexual reproduction as a particular form of concresence must have some interesting features.

I believe a connection to nature is implicitly whiteheadian, without the words. Making this change in our perception is necessary, understanding Whitehead more easily would be a secondary benefit.

That is a very core concept.

My second, unasked, question to Spryidon may go to this.

Prehension is based in feeling, more recent science identifies varieties of emotion that are shared by much of the animal kingdom. I suspect that these details would be relevant to your question.

The list of common emotions is in the first paragraph here:

https://academic.oup.com/bioscience/article/50/10/861/233998

I’ve not heard a process philosophy discussion of ‘deep learning’ as in AI. Deep learning is a more complex form of computation, and maybe also allowing for emergence in those ‘deep’ levels.

I don’t understand deep learning well enough to go further, but I suspect it is quite relevant.

I suspect that a lack of English neurotypicality is an advantage. Whithead certainly is writing to get past this but he is still writing in English.

Please don’t be afraid bring up this subject in relation to any course item.

I speculate on this question and I like the way you frame it.

I don’t have a definitive and final answer, but I do think it is safe to say that an electron has interiority and thus a subjective existence.

The questions arising for me regards experience. Does the past of an electron change it all? Does that past history effect any random measure of an electron? Does the electron ‘benefit’ from historical interior subjective experience?

Maybe, maybe not – and more likely not.

I’ll throw in an idea. Perhaps how the nature of interiority manifests matters at different levels, at the level of an electron the relation to the exterior world comes from its relation to the field, ie as a ‘wave’ identity. As complexity increases that relation to the external world is increasingly objectively ‘modeled’ inside, though the field/wave relation never completely disappears.

Does that make sense?