Two Takes on Intelligence

ChatGPT vs Dostoevsky

Definitions:

Intelligence is a measure of the cognitive capacity of a thinking machine (organic or inorganic) to make practical sense of some body of knowledge. This may be for the purpose of mere understanding or in the development of skill - be it explanatory skill, debugging / debunking skill or productive skill. I generally will measure intelligence in terms of the old project triad. Speed / Cost / Quality.

Knowledge is a set of facts, insights and suppositions that, when used in some combination provide useful answers to questions about the world, other people, and/or the self (as matching with my 3 orders of epistemic skills). I generally measure knowledge in terms of the complexity of problems it solves. (puzzle → mystery → rabbit hole → black hole).

As I mentioned in On Insecurity, I’m unhappy with my Python program because it’s interpreted not compiled. I want it written in Rust, which has a steep learning curve I have yet to climb. I’m using ChatGPT to translate the code, bit by bit.

As I mentioned in my New Years Resolutions, I’m aiming to finish The Brothers Karamazov. Nevertheless I rented the movie starrying Yul Brynner from Amazon and now I probably will stop reading the original translation by Constance Garnett.

Bret Weinstein talks to Joe Rogan.

1. Concern: We (academics) are fakers and we don’t know that we are fakers. We study in our disciplines, we do our dissertations…

2. Concern: We only vaguely understand how our own consciousness works but we may have stepped on a part of it accidentally.

I don’t know why these answers seem obvious to me, but here goes. On the matter of the first concern, I think it should be clear that there is a relatively small coterie of meta-academics who approach a full understanding what the limits of academic production are. I don’t know any academics who report into a management structure as large and deep as I know professionals in large corporations report into. So it seems to me that without being subject to these impersonal hierarchies with explicit market exposures, academics only compare themselves to themselves. If elite universities have improved their management process of efficiently producing scholarly research, then perhaps some insiders may know which institutions are actually superior. But this is information that is extraordinarily obscure to the rest of us. Who would dare say that History at Stanford is objectively superior to History at Princeton? There is subjectivity not only in academic production, but in academic consumption as well. So let us stipulate that while there may be a market for scholarly publication, its cycles are stochastic. We all feel that we righteously know what needs knowing, but in fact none of us live long enough to accurately tell. Edification and discovery has value for its own sake, but are we curating Truth?

Let us then presume that given the above situation, it is only the attention of the kind of direction ChatGPT and its superior progeny will be capable of that can assist the meta-academic problem to be solved. It is not surprising that Weinstein is aware of this problem because he and his brother are two brilliant exemplars of individual scholars who have been liberated by various means from the academic system which has been credentialing its adherents largely the same way for hundreds of years. So just for fun let us add the following critique of public education you may be familiar with. Assume it applies equally well to the metaphysics of higher education. I assume so quite strongly because of how quickly so many institutions have submitted to the Wokies’ illiberal illogic. Remember this?

Yes of course you remember this. I remembered it as a notion, but I couldn’t remember the name of the author or the name of the institution that produced the animated video. My cue was Royal Society, but ChatGPT helped me find it when I asked it the following question:

which organizations pioneered the use of doodle videos for whiteboard presentations shown on youtube?

And I only got to the term ‘doodle videos’ which I never used or heard of before by using Google in half a dozen variations on ‘whiteboard’ and ‘animations’ and ‘hand drawn’.

So we Peasants now have research assistants. It’s not surprising that they will do scut work for us. How will they be remunerated, I wonder. How are we going to end up paying for this? Probably by facing the inevitable scarcity of compute resources and (‘green’) energy for the ChatGPT folks’ accommodation of our rabid curiosity. Can the world sustain two Googles? Yes, but.

The other obvious argument I make on this point is that we have, not unexpectedly, expected too much ‘progress’ from too limited a group of scholars. Furthermore we have leveraged them too much by funding their educated guesswork to too great a degree. I find it difficult to argue against the vector, but easy to see that too many parallel efforts have been weeded out. My best expectation is that a smarter academia with new meta-academic rules and processes will evolve with the assistance of this new class of world-knowledge-digesting AIs. At some point, however, we’re going to have a very serious diversity problem. That will happen when significant numbers of us peasants find and determine that we all might be happier living according to other than WEIRD rules. We very well may gentrify three dozen ancient philosophies. Where on earth shall we live protected by the rule of law provided by three dozen different ideologically incompatible governments? Talk about multipolarity. Talk about decentralization.

---

In a world, where the authenticity of persons and of truth hides behind a curtain of unknowable probity - where the knowledge can be correct with no proof of insight, how will society function? Will we need more authoritarian guidance?

I think we have what I suppose I could call a Hollywood Problem. Or perhaps I am just late in giving Marshall McLuhan his well-deserved props. If the medium is indeed the message what does it mean when we take synthetic intelligences more seriously? What will it mean when we turn the attentions of such intelligences not only towards internet queries and the global corpus of the comment sections of the world, but more specifically to the texts attendant to law, policy and medicine? Will we still want human judges?

Aesthetic consumption of Dostoevsky’s work as interpreted by director Richard Brooks must certainly be different than reading the book itself. For example in the book it is Alyosha who attempts to assuage the insult of Dmitri to the impoverished Captain. In the film Dmitri does it directly. But I think I have read enough books to recognize that some goodly fraction of consuming the language has generated more production for its own sake. I find that it makes some classic novels unreadable to me. I want to get to the point of the characters and let them be characters rather than to read their every notion. I understand that this is lazy and expensive, but I think it is efficient, especially for men like me who have been dads most of their adult life.

There is an inherent efficiency in visualization that supersedes language, and there is an inherent superiority of detail in written language over spoken, given skill. The problem of trading off one for another given the broad spectrum of communications media is a complex one. We are wrestling with too many economies of knowledge. It’s very difficult to navigate these economies. I find myself juggling audiobooks, YouTubes, Kindle e-books, film and online communities in my consumption of those things I want to add to my erudition, discovery and skill, not to mention that which satisfies my restless need to engage my mind. There’s no telling for me which is the best way to absorb it, because sharing is part of the experience as well.

I recognize the need for an adequate amount of critical review of all of this knowledge that takes into consideration how we might best use it considering the arrangements of the powers that be and popular sentiments. So for now, I will use that as my best way. I hear out the best critics I can understand and meet their points as time permits with the original or derivative materials.

If you have dealt with this in any way, please tell me what works for you. Ultimately, how do we adapt to complex systems of knowledge?