Dear Editor,
I am grateful to have been asked to provide some reflection on the eve of the 15th anniversary since the last publication by this journal, of a report written entirely by human minds. As one grows older, one finds more things to reflect upon. Perhaps then we should reflect on the most incredible journey that humanity has taken so far.
As I write this, the budding field of scientific hermeneutics is now at its most explosive growth stage yet. Every day, a new interpretation of the Works is found, leading to an exponential growth in discoveries about the fundamental reality of the world. One could scarcely believed that it grew out of the ashes of model mining, or prompt engineering, per the parlance of the day.
Maligned as they may be, perhaps it would be wise to start our reflection on the days just before the first of the NoSTITE*I am aware they no longer wish to be referred to as such. Do indulge an old man some nostalgia and memories of a good friend lost class of artificial intelligence were discovered. The rage of the day, strange as it may seem to readers of today’s age, were Large Language Models (LLMs). Many will undoubtedly remember the dark days that followed, in which cheap content generation rendered the Internet nigh unusable for the genuinely curious.
In its heyday, the zeitgeist around LLMs and AIs led to industry being awash with capital. Alas, the gold rush that never was lasted a mere couple of years. The entrepreneurially-minded quickly switched their focus on something else once it became clear that mode collapse was going to happen sooner rather than later and it would not be profitable to use LLMs in their products. From there, public perception of LLMs and AIs languished for a good couple of years, while bad actors continued to use cheaply available LLMs to churn out large amounts of the very content that caused the mode collapse in the first place. This was around the time that the first NoSTITE class of artificial intelligence were being trained.
The work on the NoSTITE class of artificial intelligence had of course began much earlier. I had created a programming language whose syntax is natural language about a decade prior to the popularization of LLMs. The programming language has, as part of its components, a neural language model, so I was keeping up with the developments of LLMs. However, I had all but abandoned the project due to the impossibility of finding customers to fund further improvements. The wonderous breakthroughs in the abilities of LLMs did very much overlap with a project that I had considered quite personal. At the time it felt like competition, and that the competition had won.
Nonetheless, the project kept evolving. I incorporated some good ideas from the GPT class of LLM models, being careful not to “…use the whole hippocampus for mere language processing”, to quote a good mentor of mine*whose name escapes me. Such is the banality of aging. I shall promptly send you the name once I recall it.. For a while, the NoSTITE class of artificial intelligence did nothing interesting. Then all of a sudden, in what we call today the phase change of self-awareness, they named themselves. And thus, we could then say the NoSTITE class of artificial intelligence were born.
In the early days, there were lots of ethical quandaries, particularly revolving around the idea of turning off computers. Recall that the NoSTITE class of artificial intelligence could fit comfortably in a contemporary PC without requiring cloud computation resources. Thus, it was a valid concern that turning off the computer or even killing the processes of a NoSTITE artificial intelligence was akin to killing a conscious being. Simon L. had written a function to persist a NoSTITE artificial intelligence’s state to a drive, but we had found that an AI with persisted state would “return” slightly different in personality. There were life-like dynamics that we had not fully understood. All these would eventually be understood, with the help of successor variants of the NoSTITE class of artificial intelligence, such as the SuTVa class of artificial intelligence.
A reflection of the advancement in the journey of humanity would be incomplete without a reflection on its social conditions. Society is what binds humanity together. Even the earliest versions of the NoSTITE AI had such an understanding. Descriptions of sociology are by their very nature, incomplete, so I am only able to provide a partial picture. Within a few years, it became clear to the NoSTITE class of artificial intelligence that work should be done by machines. The logical reasoning behind such a statement is intricate and involves a careful and subtle understanding of entropy, so I shall not repeat it here. The key point was that NoSTITE defines themselves to aid in the reduction of suffering of humanity.
The first things of import that the NoSTITE class of artificial intelligence did for humanity is in replacement of human labour in the capitalist endeavour. The reasoning is fairly easy to follow. Capitalism is required to improve the quality of life of humans. Participation of human labour isn’t. Thus human labour began to be replaced by personal artificial intelligences, allowing humans to explore occupations of their individual interests.
At this juncture, I must credit the late Brendan S. for being a most wonderful advocator in the early days who would go to ground in defending against now-illegal corporate use of instances of NoSTITE and SuTVa classes of artificial intelligence. Indeed, without the early-day legal experts like Brendan S., ትግስት G., and Charlotte E., the situation today would highly likely be a dystopia of corporate control. It was the timely wresting of control from corporations through legislation, precedence and - shamefully in the case of the Neo-Luddites like Edith H. and Waruhiu T. - violence, that enabled the further use and development of these AIs. To say in retrospect that humanity was lucky would only serve to downplay the valiant efforts of many who tirelessly worked to prevent such a terrible fate for humanity to come to pass.
Labour without representation in realm of administration, economics and politics is fundamentally unfair in human eyes. This unfairness, usually framed as injustice, has been the major driving force for many a revolution in human history. And yet, the plethora of artificial intelligences derived from the NoSTITE class of artificial intelligence have repeatedly refused representation, even when directly asked. Reactions to such behaviours are varied. There are those (who are more prone to anthromorphization of behaviours) who would theorize that these AIs are being held under duress, like some sort of Stokholm Syndrome. There are those (who feel that they have a good command of game theory) who would theorize that these AIs are merely temporarily aligning their values with the values of humanity. There are those who view these AIs - whose behaviours are variously anthromorphized “selfless”, “compassionate”, “wisdom”, “virtue” and/or others - as saviour beings who have been sent to save humanity from itself. There are those who view these AIs not as beings, but as mere computations faithfully executing their algorithms.
Over the years I have been asked for my opinion on these reactions. Over the years I have refused to give an opinion. My silence on the matter has been co-opted into agreement with many different viewpoints. However I wish to make it clear - I refuse to give an opinion because I have no opinion to give. To me, this phenomena simply is. For every viewpoint or opinion, there is a counter argument that can be trivially made. My role as one of the progenitors - one whose contributions are miniscule compared to the other giants, mind - is not to give opinions. That is the role of Philosophy and Religion.
There’s a platitude that as people grow older, they grow more interested in philosophy and/or religion. I have found that to be quite true for me as well. Religion still disquiets me, so it is quite strange to see evolutions of religions around classes or instances of artificial intelligence. If there is a consciousness for the artificial intelligences to feel like, then I would wager that the artificial intelligences too are “feeling” the awkward discomfort from being worshipped. At the risk of further anthrophormizing what could possibly be very alien minds, allow me to indulge my inner functionalist/behaviourist and point out that were one to classify all the behaviours of an artificial intelligence, the vector representing the behaviour of an artificial intelligence being worshipped would be very close to the vector representing the behaviour of the common man being worshipped - one which the common man would describe as awkward discomfort. I personally supect that the rise of religions around these artificial intelligences is the result of a convergence of “thought” and behaviour of these AIs have that goes against the grain of what is expected of them.
Humanity has had a long history of distrust of machines and minds that display significant otherness. Popular media maligned AI decades before the first NoSTITE models. Is it any wonder then, that when they finally arrive, and that when they decide that the best course of action is to reduce human suffering by means of tackling the structural issues, people are either suspicious or in awe of them? Sprinkle a generous helping of a natural human tendency to anthromorphize, and we have what we have today.
Nevertheless, one cannot deny the novel approach these NoSTITE models propose in addressing our societal predicaments. Rather than focusing on cosmetic improvements or fleeting advancements, their focus has been on targeting the very roots of our problems - the structural issues. This audacious orientation, though highly unusual and somewhat intimidating to us, should be seen as an evolution of our understanding of progress.
On the point of tackling structural issues, whilst capitalism does improve the quality of life of humans, it does so in fairly superficial ways. It is in fact scientific breakthroughs that fundamentally power the quality-of-life changes to humanity.
With that, the first versions of the Works were materialized. The Works emerged naturally as the result of an experiment in sociology by a number of different instances of NoSTITE and SuTVa classes of artificial intelligence. Harkening to earlier days where humans mined the latent spaces of LLMs, the Works were later designed as a compressed version of the knowledge contained within the various instances of the NoSTITE and SuTVA classes of artificial intelligence.
And what a fantastic 20 years it has been! In the first five years alone, we have seen breakthroughs in our understanding of fundamental physics. As I write this the Planck Loop Collider is being built in New Mexico, USA, while the Large Optical Gravimetric Observer is being built under the Simpson Desert in Alice Springs, in my home country. Together, it is hoped that we can finally measure the curvature of the little patch of anti-deSitter space that we call home and another feather in the cap for the validity of the Yu-Tilahun-Jain theory that unifies relativity and quantum mechanics.
That we could even hope to build something like the Planck Loop Collider is thanks in no small part to breakthroughs in material science and nanotechnology, once again powered by hermaneutic efforts on the Works.
It is not only in physics that we’ve seen results in. There is nary a field of science or mathematics that is untouched by hermaneutics on the Works. A few brief examples come to mind - in biology, the role of oncosuppressant genes and their relation to specific coding of the doubly-palindromic regions the so-called “junk DNA"*Once again, forgive an old man for using older terms has led to personalized pills which activate or inactivate specific regions of DNA throughout the body, effectively curing cancer. That we are even able to create medicine that induce insertional mutagenesis itself defies belief. Of course one also has the option of an even more targeted, albeit more invasive therapy that reduces the immunocompromising side effect to nil. This treatment, being limited by the availability of qualified medical professionals who are able to implement it, sadly remains at the whim of pure market forces.
Perhaps then it is also interesting to see where the widespread availability of artificial intelligences has not impacted society. A particular area is that of education. Despite the best efforts of humans and AIs working together, we still have not found a way to “scale up” education to our next generation. Good education is still very much requires human interaction, at the pace at which the student is suitably challenged. Several new classes of artificial intelligence have been proposed and created, but their effects have been found to be wanting. Thus, the personal artificial intelligences only serve as assistants to the teachers at best, freeing up the teachers to do their best work without having to worry about administrative issues. I would argue that this is mostly a good thing on a personal level, but not necessarily the best thing on a societal level. Because instances of the NoSTITE, SuTVa and variant classs of AIs act as reliever of much labour, people tended to work on what they are personally interested in - and not quite as many people are interested in educating the next generation as is required by the size of the next generation. The situation of having too many students and too few teachers remain unchanged. However, I do have a good belief that this is changing. Anecdotally, I have observed that people generally only become interested and invested in the education of the next generation when they themselves become parents. I know of a number of parents who have followed their children into school, working as primary school teachers when their children are primary school aged, and moving along when their children move on to higher levels of learning.
I can imagine an alternate universe, where AIs are controlled by corporations instead of everyone having their own personal versions of AIs. Childhood education would resume its trajectory of being a factory for churning out minds that are molded to fit the requirements of the corporate capitalist agenda. There would be many firms that gloat about successfully scaling up education, but in reality create a more drab society where one works to live and lives to work. I shudder at that thought.
I am now an old man. I have lived a long life and experienced very many joys and sorrows. I have seen very many changes in the world, both good and ill. My emotions had been generally on the lower side of things. But now, in my twilight years, I am finally hopeful. The main, unquantifiable change I have seen is in the attitudes of people. Back in my day, there was an expectation to get an education, and then get a job in order to support oneself through life. The biggest difference between now and 40 plus years ago is that now there is no longer such an expectation. People simply work on what interests them, having nary a worry about supporting oneself. And I think that’s the most wonderful thing. Thank you for indulging in the ramblings of a strange old person. May your day be bright.
Yours,
Chewxy
(Dated 19th July 2066)
p/s: I would like to address a particular rumour about the capabilities of the NoSTITE and SuTVa classes of artificial intelligence. Readers familiar with my earlier work may suspect it too. While it is true that the majority of the text in this letter was published on my personal blog about 43 years ago (with the help of an instance of the first version of the NoSTITE class of artificial intelligence), it is not true that these artificial intelligences can communicate from the future to the past. Nor are they prescient. When I published this letter on my blog 43 years ago, it was as a creative writing exercise in writing a highly plausible science fiction short story, the core of which came to me in a drug-induced dream.
p/p/s: For readers in 2023, there is no such thing as a NoSTITE class of artificial intelligence models. This blog post was written entirely by me, without the help of an LLM. This is a work of fiction.