Sunday musings…2/26/2023
1 Flies. “Like flies on shit.” At the moment I am at a horse barn, my GPS coordinates for at least the morning for a week.
Do the math.
2 Trace. A beaten path or small road; a track. More in a moment.
3 Mews. A row or street of houses or apartments that have been converted from stables or built to look like former stables.”An eighteenth-century mews”. Wait for it…
4 Road. Or street. What everyone in North America calls a thoroughfare. Everyone, that is, except the supremely affected class in Wellington Florida, winter home to fancy horses and the people who fancy that owning said horseflesh makes them, you know, fancy. Don’t they remember that a classic eighteenth century mews in the Uk was literally awash in shite?
No matter how fancy those horses are, or how fanciful you may believe naming your streets with names that conjure up images of a bucolic countryside and yourself landed gentry, cruising the various traces and mews in town, the lineage of these words includes bountiful amounts of horse shit.
5 AI. There is a new player, a new candidate for the force behind the next major disruption along the timeline of human intellectual evolution. Artificial Intelligence (AI) in the guise of pseudo-autonomous bots is being heralded as the vanguard of this next major step. As an aside, I hesitate to call it a “leap” given that the “intelligence” involved isn’t really human. As a matter of fact, the computer scientists and engineers who have developed the “machine learning” that has spawned this phenomenon admit that they really have no idea how it is that these AI programs work.
Imagine that. You’ve created something which is very likely the next step along the timeline that began with the drawings of cavemen, on to papyrus and the written word, through the invention of the printing press, radio, and television, and most recently the internet (and specifically internet search which makes almost all knowledge findable), and you don’t know how it works. All of these steps brought about the liberation of information, each subsequent one increasing the access to information to a greater and larger slice of humanity. Indeed, the printing press likely did more to topple the prevailing hierarchies of the day than anything before or since. Anyone living along a trace who could read now had the same access as all of the lords and ladies living above those mews.
Here is what I wrote about the internet’s effect, further driving down the barriers to information, and what this has done to the intellectual processes of a typical human, as I mused about why I muse:
“Why do I write? Why do I sit down and use time that could otherwise be put to use in the gym, or in the office, or even just hanging with the Man Cub? As a long-standing lover of language I am always on the lookout for the best vocabulary to explain concepts I sometimes struggle with. “Offloading” is a term that is used in this case to describe what it is that humans do with information that they do not need to keep on hand in “useful memory” space.
This is what I do with ideas when my “wetware” memory is full. This is hardly new. Indeed, the sturm und drang associated with the mega-trends in education, etc. associated with our massive information/recall apparatus that is the internet actually has its origin in the Greek era of Socrates and the transition from an oral tradition to one in which teachings were written. (HT to Frank Wilczek). Prominent adherents to the oral tradition such as Socrates and Simonides argued forcefully that the advent of the written transfer of information would weaken the mind and produce an inferior type of intelligence. In a fascinating and delicious ironic twist, all we know of either of these men we know because someone else wrote down what they recalled hearing.
In my day job we are still encased in a paradigm in which information is transferred from teacher to student and then tested to see if that information has been committed to memory. Imagine, with the explosion of data now available in the world of medicine we test (and test, and test…) both new doctors and established ones to see if they remember a certain percentage of facts, regardless of how often those facts come into play in the act of practicing medicine. A CrossFit analogy is to test a trainer on the precise moment that the obturator engages in the deadlift. One neither needs to know this to teach the deadlift, nor does one need to have memorized this in order to have it on hand in the gym. So, too, in medicine.
Please don’t get me wrong, I still enjoy knowing a bunch of stuff and being able to call up that stuff without needing to use my Google-Fu. The reality is that we have made a move from memory in written form to memory in digital form that is just as profound and disruptive as that from oral to written. We have only to remember where it is we have stored our memories, our books and our music and our musings.
And our passwords. We still need to remember our passwords.”
What, then, of AI? It surely seems as if it’s more than a simple repository of information or a new way to share that information. Or maybe, not.
As I understand it, the AI that we are seeing in ChatGPT does little more than mine our existing cache of knowledge and then regurgitate it in a way that mimics what we humans do when we are in possession of a tiny slice of that knowledge. More that that, ChatGPT is capable of discerning the patterns of an INDIVIDUAL human, then to respond to a prompt in a manner which is consistent with how that human, or “kind of” human might respond. Want a picture in the manner that one would expect from Salvador Dahli? Sure. Piece of cake. Here you go. How about a Shakespearean sonnet, pegged to any time in recorded history? What the heck…make it in the style of Ben Affleck and give ChatGPT a chance to woo JLo. Again, piece of cake.
But is that a form of intelligence? A kind of “mind” capable of something truly original? Or even an original thought or concept which is derivational from something prior, as is much of what we consider conceptual development? I’m saying “no”.
The potential for this particular form of faux AI is that is has provides the possibility of relieving the role of humans from a very large percentage of tasks presently requiring “wetware”. Customer service. The process of registering for pretty much any service imaginable (e.g. medical). It can provide a probabilistic answer to an almost immeasurable number of tasks. For example, in my world of medicine this type of AI could be used to force the use of algorithmic protocols, albeit using the best of knowledge available as well as probabilistic (Baye’s Theorem) methods in coming to suggested diagnoses and treatments. After all, it’s just math, and in reality all AI in its present form does is a kind of math.
Shakespeare’s genius has been reduced to the mendacity of the binary.
Is this a bad thing? Ah, here lies the rub. In a world where more and more of the decisions that are being made for us have been reduced to math, the technology behind ChatGPT and the like gives ever more power to those who would profit most through the mathification of the individual. Would this work for the majority of humans a majority of the time? Of course. Common things are common, and this type of “intelligence” is designed to parse the common.
It is in the soft surroundings that we see the danger in this next, latest great “advance” in the liberation of information, precisely because this is the first iteration in this inexorable progression in which the information is not actually liberated at all. It is, in essence, captured through the process. It is a regression to the mean, albeit many precisely defined means (Dahli, Shakespeare, Affleck, et al). After all, what each prior advance in our liberation of knowledge, each leap in the ease with which a greater percentage of humans could acquire that knowledge has done is make the next truly human genius leap accessible to more humans.
This is NOT what the faux AI of ChatGPT is doing.
A true intellect, one that is alive, intelligent, creates something new. Truly new. Something that has heretofore not existed. Cell theory. Quantum physics. Gravity. Relativity. Heck, Baye’s Theorem. Nothing that we know about AI as it now stands leads us to believe that it will solve the essential problems that lie in the pursuit of energy from fusion rather than fission. The work being done to decode the unbuilding blocks of aging is being done between the ears of living humans, not AI. The best, and worst, that we can say about the “AI” implied by the likes of ChatGPT and Bing is that it has the possibility of relieving humans from the menial, the mundane, and most damningly, the meaningful.
What will we become if we cannot, or do not interact with one another when it is meaningful? See “Ex Machina” for a particularly apt example.
In the end, or at least this beginning of the end, AI as we presently experience it is nothing more than the wrote recital of an admittedly more complex rendition of the present knowledge base of humankind. In a sense it curates the massive volume of knowledge sitting behind a query and presents it us in a manner that makes it digestible. The written word allowed us to preserve the thoughts of men such as Aristotle. The printing press made it possible for any human to read to have access to Socrates, et al. A combination of the world-wide web and the marvel that is the search engine essentially removed the need to memorize Socrates’ words, while at the same time expanded access to him, and all of his ilk, to a nearly every demographic on the planet. ChatGPT and others of its ilk can tap into the vast amount of information from, and about Socrates, curate and cull it, and then present us with a cohesive, one-bite whole.
But is that intelligence? iI is a singularly human trait to be able to conjure the truly original. As fascinating, and admittedly entertaining as it is, the AI we see demonstrated by ChatGPT, Bing, and the like is not intelligence if we define intelligence as having the capacity to create something truly original. Might it be convenient? For sure. Why not? Is it, as it is now, an unabashed and total positive? Of course not. The internet/WWW/browser universe has diminished our human intellect insofar as we as a species tend not to commit too very much to our “wetware” memory any longer. I can’t help but think that will somehow come back to haunt us, in much the same way that allowing the juvenile, barely above trivial abilities of ChatGPT and the like to take over our mundane tasks. What if we conflate the mundane (fixing a plane reservation) with something as meaningful as falling in love (vs. the mathematics of having all of our desires and expectations anticipated and fulfilled a la Ex Machina)?
In the end this is no more than another evolutionary step, and in my opinion a rather small one from a technical standpoint. It is real, for sure, and it is meaningful. But printing press meaningful? Internet meaningful? Not yet, I don’t think. Not at the level of the individual, the almost sacred level of the enlightened world. We risk demeaning our very humanity if we delegate too very much to an “intellect” which is still nothing but 0’s and 1’s. It’s nothing more than math, after all. We have used math in ever more ways over the millennia to improve the life of humans while holding tight to the non-math that defines what it is to be, you know, human.
Like feeling empathy for a patient. Or falling in love.
What passes as AI at the moment is nothing more than a tool, albeit one that will need tight leash of unbreakable rules (see Asimov: The Three Rules of Robotics). In the end intelligence, real intelligence, intelligence that moves us ahead as a species in a meaningful was as did the written word, the printing press, and the access to both afforded by the internet, comes from a place not yet visited by the artifice of ChatGPT and the like. We will need to guard against allowing such an immature intellect to provide much of anything at all in our lives, in my opinion, other than the curation of content. After all, we know not who it was to whom Shakespeare wrote his most loving sonnets, but we know that he did, indeed, love whoever it was. A new, truly real intellect will need to do more than simply create a sonnet a la Affleck as Shakespeare, professing love for Jennifer Lopez. Humans are capable of love. Whatever comes next must not only be able to speak lovingly of JLo.
To be real, to be intelligent, whatever comes next must truly be able to love Jennifer Lopez.
It’ll be me, in person, and I’ll see you next week…
This entry was posted on Sunday, February 26th, 2023 at 9:47 pm and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.