Introduction

The establishment of this substack came after much deliberation, both internally and in discussion with friends and colleagues. It would then be appropriate to present the historical background and the primary motivations and goals associated with the many articles I have already in mind to write, as well as the expected indefinite output thereafter.

Like many initiatives, a mid-life crisis set the backdrop. I have been working quite some time as a research mathematician in numerical optimization. For the most part it seemed as if I had published loads of analysis and simulations of optimization procedures that never ended up actually optimizing anything in real life. But no matter – I knew there was structure to the broad enterprise of academic research. More general results are later extended and collated by others in the field to make software and techniques that actually result in some optimization of a process or learning procedure that would, hopefully, serve some positive social utility. That is, academic research was a grand cooperative enterprise, where specialization would eventually be connected by the structure of dissemination and knowledge network diffusion.

This illusion came crashing down, however, a year ago. After funding for general research was going to become scarce in the department I am in, I helped an individual write some grant proposals for applied research, with the topics such that I was figuring there’d be some natural obvious things that would be done on these projects with some marginal but positively beneficial utility. However, after the projects began, at some point I found out that the algorithms being developed in one were AI procedures performing major economic interventions that would, if implemented, be highly likely to significantly increase poverty and mortality rates in Europe. With another project, it then appeared the PI was focusing all of his attention of personal interests irrelevant to actual project data.

The details of these will be the subject of later articles. However, the experience was psychologically turbulent and utterly Kafkaesque. I was constantly baffled by what was happening. It seemed if as if nothing he was doing made any sense, while he was mean-spirited in doggedly pursuing it.

So as Einstein was fond of saying, the source of the problem lies at a higher order of abstraction than how the problem presented itself. The evidence simply was not there for the low resolution explanation of either complete ignorance or unfettered evil being the source of what I encountered. The problems were more systemic. Indeed, implicit support from the government institutions funding the project underscore this fact. And while I cannot read minds I will never know the exact specific details, but ultimately discussions with friends and later comprehensive revisiting of many books I had read before, and reading of new books as well as various substacks, led me to conclude that he is rather a middle managing NPC in grand systemic processes that establish just the right incentives that inevitably lead to comprehensive and ubiquitous dysfunction within research.

One may claim such an identification, of some sort of rationalizing self-delusion, is akin to accusation of stupidity. I bet to differ, rather, motivated reasoning and rationalized self delusion appears among some of the greatest in cognitive ability. For one example, Eric Hobsbawm was by all accounts a brilliant historian, who, however, never dropped his support for the Communist International even after repeated confirmation of Stalin’s horrors. Bertrand Russell, one of the greatest logicians and philosophers of all time, urged the UK government to de-arm in the 1930s in the name of pacifism. This sort of ivory tower utopian virtue signaling could have had obviously disastrous consequences if the government took him seriously.

And so the details of these two projects and the Kafkaesque experience of confronting them while being baffled as to even why, will be the subject of future articles, but what is far more important is what I learned and the way I would begin to think about things, which radically transformed over a year, in a period of rather heavy psychological turbulence. I learned much about the world around us and made important decisions regarding priorities and values. I will describe the peak of my efforts for human welfare positive research with Agroforestry innovation in East Africa towards the end of this as well as in future articles.

In my quest to, upon facing the events with this PI, figure out WTF is going on, I dived into piles of reading and contemplation. My deep retreat into history, philosophy, economics, etc. was nostalgic. For I had a phase in my youth, due to autistic lonely childhood, of readings a great collection of books on these topics, as well as long discussions on an online intellectual discussion forum at the time. And ultimately, the circumstances leading to the tragedies smacking me out of stupor and into class consciousness as a member of the intelligentsia, were overdetermined – there are three layers of grand societal dynamics and patterns, all three of which have a number of supporting scholarly literature. It was almost as if how could I be so blind.

Crisis in Academic Research

I started reading about contemporary phenomena in aggregate with respect to scientific research and came upon quite some starting findings.

Why Most Published Findings are False

The most salient news is the replication crisis. Fast swathes of published research, initially with social science, with more than 2/3 failure to replicate rates for some subfields, the worst being social psychology, to even medical research, wherein estimates hover up to 50%. This is caused directly by several things: simple type one error; p-hacking, publication bias, and other statistical anomalies; and downright fraud. Many classical findings that were canonical knowledge in textbooks, such as ego depletion, were discredited from a repeated failure to replicate. Some of the worst outcomes involved medical research that proved fraudulent that was used to perform surgeries, and this resulted in significant excess mortality. As far as my field, applied math, replication is labor intensive and so only small sample sizes have been used, but I’ve seen it estimated at 20% failure to replicate.

A bit of a whistle-blow: the most groundbreaking advances in Machine Learning and AI are released in several conferences, including NeurIPs, CVPR, ICLR, AAAI, and others. All of these examples I’ve published in, as well as reviewed in, and so I’ve witnessed the process of peer review. A reviewer gets a few weeks to perform 2-5 reviews, and make a decision and commentary. The main part of the paper is usually 8-10 pages, with an unlimited appendix, which, however, the reviewers are not required to read. The proofs of all the theoretical results are in the appendix. So, as you can imagine, many of the published works have theoretical resulted with proofs that haven’t been peer-reviewed. There are exceptions, my recent NeurIPs publication has an extended theoretical section in the appendix after a back and forth with a reviewer. But I have heard that a nice exercise for a grad student is to try to find mistakes in random ML/AI conference papers. To give proper respect, however, the review process in the journals in this field, JMLR, ML, IEEE TNLS, etc. have been notably more professional and thorough.

Moreover, the same computer vision and language model datasets are often used to demonstrate new models and algorithms, and it’s been conjectured that a lot of improvements have actually just been greater overtraining. In recent times, these conferences’ publications are becoming increasingly from BigTech, who just throw GPUs at a problem.

Now, consider the ultimate economic purpose of STEM research: to develop fundamental science that results in new technology that defines a new innovation in the economy, raising marginal productivity, and thus long term prosperity. (I suppose social science is for “having an informed population on the human condition”, or training the propaganda apparatus, depending on how you look at it) Now, has it been doing this?

Somehow no:

Papers and Patents are Becoming Less Disruptive Over Time

Marginal Productivity Gain Relative to Researchers and Spending has Sharply Declined

It seems like there is an increasing stagnation as far as the innovation generated by STEM research. Even with additional resources in funding and people, novel technology is still stagnating. This clearly fit my experiences with the applied projects, and so what is causing this? Various theories exist, for instance the possibility that all the low hanging fruit were found. There are also various problematic incentives that exist in the system. My personal belief is that it seems that increased highly specialized knowledge has a flat marginal productivity gain. Whereas there is a dearth of considerations such as faithful modeling and thorough and transparent uncertainty quantification to real data and means of increasing clarity within a broader context. I do see an important and critical step is to develop systematized interdisciplinary knowledge, wherein different scientific language games and models are integrated into cohesive wholes. I will go into this in more detail later.

The Symbolic Capital Bubble

Symbolic Capital is that which is purely in the form of ideas, such as text or video presentations. Coinciding with the decline in manufacturing and increase in services, more and more of the economy as well as employment involved with symbolic capital. This included, after deregulation, an explosion in the media industry, an expansion in higher education, an increase in corporate concentration and thus bureaucracy and branding work.

This can be described through a number of perspectives:

Schumpeter, in Capitalism, Socialism, and Democracy, presented a prognosis that with the increasing prosperity in the west, ultimately an intellectual class will swell that is hostile to free markets. At the same time, markets will increasingly concentrate over time while the burgeoning monopolistic capitalism will become intimately intertwined with the state. This new corporatist economic regime will stifle innovation, which will ultimately result in stagnating productivity and living standards. The process of innovation, creative destruction, requires flexible markets with slack demand that can direct investment flows from sclerotic to novel technologies. Oligopolies with extensive government organizational ties will ensure these flows remain in the hands of established companies and elites. It can be quickly seen that all of his prognoses became true empirically, from monopolistic concentration to stagnating productivity, and the grand expansion in the size of federal employment and spending over the 20^th and beginning of this century.

Jean Baurdrillard is a noted French philosopher who studied and presented comprehensive accounts of how signs, rather than tangible objects and experience, have become front and center in the contemporary economy. In Critique of the Political Economy of the Sign he echoes and goes beyond Thorstein Veblen’s work on the consumption patterns of the luxury classes and how their prices follows symbolic memetic dynamics. Similar to the Frankfurt School, he departs from Marx in denying dialectical materialism and noting the central role of mass culture for the management of the PR of modern capitalism. He notes how the social affiliation and status identification of consumption becomes the primary consideration over the good’s practical utility and benefit to the consumer. His most famous work, Simulacra and Simulation, is a well respected depiction of contemporary life as completely disjoint from the real sense of day to day experience, and is a dizzying array of cooperation and competition in branding, hegemonic and subculture narratives, value identification with consumer choices, and vast swathes of employment working with purely Symbolic Capital. That is, rather than tangible goods that embody physical objects or a direct service to a human being, more and more of the substrate of economic production is written text and audiovisual media. Considering Derrida’s observation of signifiers pointing to signifiers etc. being the structure of meaning, it can be seen how this can swell into increasingly divorced-from-reality theory-crafting by the Professional Managerial Class, the expanding social stratum working in government, academia, media and corporate(marketing, etc.). This class’s interests being divorced from the working class, they hold beliefs that are starkly different from the general population of the west, let alone the world. Furthermore, its brand of politics of statements of policy as signalling has created the environment for the rise in Authoritarian Left politics.

There are many more, but I conclude the literature review with David Graeber and his Bullshit Jobs. He observed that there is a large portion of the employed who are performing tasks that do not provide any value-added, that is, any direct change to a good or service, and otherwise don’t contribute to economic output. These fill up the administrative bloat of corporations, arising from various regulations, and monopolistic practices. His colorful taxonomy includes Flunkies, who just add an aesthetic of high status to the boss such as receptionists and doormen, Goons who play tactics of optics for the company such as lobbyists and marketing, Box Tickers, who just keep track of things, and Taskmasters who just delegate work.

One of the serious problems, of course, is that if these jobs don’t add to aggregate productivity, they are a drain on the economy. Indeed, consider the often reported statistic that real worker’s wages have been in stagnation for 50 years...but productivity has gone up with respect to their work. The fact is, because of the Baumol effect, the flat productivity of the symbolic economy brings down overall gain in wages from gain in productivity, and so income growth for everyone employed in the firms, including the more productive working class, is anemic.

Finally, there is also a political element, as recent times have been characterized by the rise of Authoritarian Left ideology. Observe an interesting fact: in the last 10 years the working class in the west has departed the left and went into the arms of the New Right. The intelligentsia, once stalwart allies of the working class, has lost this bond. What happened? A lot of the more Authoritarian Left elements, coming from (a rather aggressive reading of, which is a topic for a future article) the Frankfurt School of Heterodox Marxism, seemed to take over academia, especially in the social sciences, following the movements of the 1960s in the west. However, ultimately, it happened exactly how Bakunin warned in his break from Marx at the First International: a workers’ movement that is statist, rather than libertarian, in its foundation, will only lead to a horrifyingly oppressive dictatorship. His premonitions being vindicated throughout the 20^th century only go to show how natural it is for a vanguard class to claim to represent the oppressed, while setting up networks of patronage and power serving themselves. Such is the situation in the modern day, with celebrity activists and entire fields of nonsense.

Grievance Studies Affair As a nice illustration, consider the project known as the Grievance Studies Affair. Here, a number of hoax articles deliberately composed of nonsense were written and submitted to several journals associated with some of the more activist-heavy critical theory practicing fields in the social sciences and humanities. 4 articles were published, 3 accepted, and 7 under review. Among the published articles included an investigation of dogs engaging in “rape culture” at the dog park and a framework for “Fat Bodybuilding”.

Lest you think STEM is immune from self-serving nonsense: while I’ve heard mixed things from physicists about Sabine Hossenfelder, this is a pretty damning smoking gun.

A Terminal Phase in the Decline of Modern Civilization

The two preceding phenomena coincide with several others, that appear to form a consistent set throughout history. Specifically, beyond stagnation in science and bureaucratic bloat, the modern day features a decline in religion and an increase in individualism and hedonistic pursuits, steep decline in population mental health, increasing number of people dropping out from modern life economically and socially, decreasing fertility and coupling rates, political polarization and radicalization, and resource depletion and environmental concerns. All of these have appeared as indications of a civilization that has entered the terminal phase that will proceed towards decline. Economically, it is already expected that Gen X will have the peak income and wealth, with millenials the first generation worse off than their parents. All of these civilizational aggregate patterns have appeared many times in history. This includes the Roman Empire, the Byzantine Empire, the Ottoman Empire, the Babylonian Empire, etc.

Literature on this phenomenon goes back as far as Spengler Decline of the West written in two parts, published in 1918 and 1922. He described civilizations as having cyclical expansion and decline patterns and that in a century this would affect the west. Arnold Toynbee is the most respected prolific scholar on the study of civilization, with an extensive A Study of History spanning 26 civilizations’ history of ascent and decline. It has been observed that many of the contemporary phenomena mentioned above coincide with characteristic features of decline in his descriptions.

A systems theory perspective is given by Joseph Tainter The Collapse of Complex Societies. In this work he describes a society as composed of systems that perform functions and solve problems. However, the systems themselves introduce some problem and additional complexity is continuously added to the social system. At some point, the marginal utility of additional complexity flattens, and the system is overburdened with the requirements of maintenance. Ultimately, the society collapses, and declines and/or is taken over by a competing civilization.

Finally, End Times by Peter Turchin involves a creative use of quantitative simulation and modeling as well as empirical information. The author studies the dynamics of different classes in society, and, in particular, analyzes how a process of elite overproduction can yield stagnation, instability, and decline. This process is reflected in the Russian revolution, among others, with the sons of landowners, after the serfs were freed, being in oversupply in the cities relative to the total number of government administration and civil service job. That 48% of college graduates are underemployed in the USA isn’t a coincidental feature of the recent polarization/radicalization.

A biological perspective is given by the mouse utopia experiment. In order to see what abundance of resources within an overcrowded space would, Calhoun famously subjected groups of rats to such circumstances. Their behavior, especially social and mating behavior, became strange and pathological.

Indeed as a species, biologically we are meant to only survive and replicate, and there is no essential reason that our basic drives and motivations should be ones towards ever greater progress. Rather, this the great expansion of human achievement and prosperity over the last several hundred years is a fortunate accident of the agglomeration of advantageous geography and resources together with the historical development of the right institutions in the developed west, which spread, to various degrees and with varying success, to other parts of the world. As it is, the challenges facing research is but one of the problems wherein humanity is at risk of general stagnation and secular decline. Moreover, unfortunately the developed world’s economic stagnation is contemporaneous with the possibility of many developing countries never escaping the middle income trap or worse. Of course, one of the challenges to any forecast for society is humanity’s ever unpredictable praxis, and this is not inevitable relative to the grand potential of human will, but it is a likely risk at this juncture.

Consequences of Systemic Dysfunction in the Knowledge Economy

So there are three different macro multiscale phenomena that are synergistically taking place, explaining the events that happened to me. The harmony in this trilevel narrative is clear: when there is so much symbolic capital it becomes a tangled web of signifiers. As there is no reality check on research work theory crafting can give rise to systemic dysfunctional patterns, and even negative consequences on the real world to zero risk for the researcher. Such an environment provides ample room for status games of virtue signalling that leads to an abundance of luxury beliefs, that is, policies and values that provide minimal risks for the intellectual upper middle class while resulting in serious problems for the working class.

In discussing these matters with colleagues, a common preference is not to actively concern oneself with such matters, and just focus on one’s narrow sliver of research. That’s certainly understandable. However, in aggregate, it presents a freerider collective action problem. If no one stands for declaring truths when necessary for science’s good faith progress towards human flourishing, then there are certain consequences.

Many academic people are troubled by the rise Trump, Le Pen, the Afd, Meloni, Wilders, the Swedish Democrats etc. etc. (clearly this is a trend) and yet their popularity was caused by, primarily, this very multi-phenomenon. Moreover, there is much concern for the public’s trust in science when it comes to messages on climate and vaccines. Yet if there is good reason to not have trust in science broadly, can you really blame them? If academics have comfortable working conditions with an upper middle class salary together with social prestige and status, they should be expected to contribute to society in a positive way, and the clear dereliction of civic duty is going to have consequences.

Indeed of working class people that I know, they do report that animosity and contempt towards the intellectual class is ubiquitous among workers in the present day. And unfortunately, if the research we are doing contributes to our own symbolic capital with little positive effect on the world, let alone the humanitarian catastrophe that would happen if the first project I mentioned is implemented at scale, then the working class has every right to hate us. And this has consequences: as I write, academics in the US are in panic over a recent order by Trump to cap National Science Foundation indirect overheads at a rate much lower than the average overhead rate for grants, leaving a potential shortfall of funding for University infrastructure and administrative staff. And so elections have consequences, and while I am writing post-portem, don’t say that the academe wasn’t warned.

While I will write some about political matters, and more specifically political matters when it comes to science, it is not the focus, rather, the blog will be a broader attempt to bring a message of greater epistemic virtue and the faithful modeling of reality to the intelligentsia. I will cover the practice and method of applied mathematics and its application to science and innovation itself. In addition I will pontificate on matters related to the higher level systemic dynamics of modern civilization. That is, the contemporary anthropological and sociological trends associated with decadent dynamics, the patterns of economic dysfunction across the world, philosophy in general and as applied to phenomena in the present day. Considerations of various non-solutions as well as solutions and/or tradeoffs to realistically consider (when there are no nice Pareto optimums) will be presented. An attempt for a meta on all these forms of inquiry, that is cross-disciplinary synergistic integration of concepts, the mechanisms of emergence, and other fun stuff will arise, too. Finally, I am officially involved in projects of contemporary social interest, namely energy optimization as well as AI Alignment, and so will provide broad analysis assisted by my technical expertise.

Motivations

I believe in particular that the marginal productivity of narrow specialized scientific research is essentially flat. Every field has a deluge of ever epsilon-extending developments in theory that have never been applied in real world contexts, nor even considered in interdisciplinary research. Both anecdotally and reported in the literature , interdisciplinary work is more difficult to publish, with different communities having different standards and expected practices and themes in scientific discourse. Fundamental challenges presenting many research questions become apparent on considering modeling, data-fitting, controlling, and other quantitative operations with the multi-domain system. The expected diffusion of knowledge towards innovation requires significant methodological lubrication.

At this junction a primary motivation becomes apparent: I wish to have a venue to broadly express interdisciplinary intellectual musings. Not mature and sophisticated results for a journal, and at technical understanding that, broadly speaking, individuals who are well read across fields would be able to understand and appreciate my ramblings. I’m sure my preferred topics, points of view, priorities, etc. are not everyone’s cup of tea. However, I’m sure there are weirdos out there who could enjoy some of my intellectual pontifications. And so in a way this could serve as to help me find my people. For you see, while I was an avid reader through high school and college, Then I expected but never experienced the “Vienna Circle” like high intellectualism across the expanse of human knowledge I had a longing for in going to a (high quality) university, then later graduate school and academic departments. But in High School I was part of an intellectual online discussion forum, which was great, and perhaps I can find and build some little community in the digital world today.

I also wanted to use it as a reference, because during communication live or by email and some related topic or question (to the topics of the upcoming articles) is asked, I can share the article that describes my thorough perspective on the topic. Often difficult and complex topics simply require a long exposition fit for nuanced analysis, such as substack provides..

Philosophically, I want to establish the proper formalization of Applied Math as truly applied to grounded reality, rather than just in the sense of research within some collection of specific fields considered to be applied math (Probability, PDEs, etc.). This blog will discuss ontologies and model representations in the context of real world circumstances, and what can be inferred from such an investigation. But it will broadly span Philosophy, namely Epistemology and the Philosophy of Science, AI, Economics, History, Sociology, and other topics, darting across disciplines.

I have a passion project in my research that brings me much excitement. Working with Kenyans and Moroccans, I am studying and working on Agroforestry. Agroforesry is the practice of combining trees with crops, which can provide medicinal foliage, firewood, fodder for livestock and other products, as well as improving the soil quality. The technology is green and sustainable, trees even uptake carbon dioxide, and has a very low capital cost. With the assortment of additional products and synergistic benefit to the land, the economic revenue of the smallholder farmers and pastoralists can achieve up to double. I am working on research as far as various computational mathematics, statistical, and AI techniques for Agroforestry in collaboration with a number of universities, forestry and other institutes, and companies in the two countries on matters such as modeling tree metabolism, statistical analysis of Agroforestry practice, and drug discovery from medicinal plants

for which we are currently working on applying various AI and machine learning algorithms for.

As far as maximinning utility or any other social welfare metric (such as my favorite of Amartya Sen’s capability vectors), I think it’s clear that AI-based innovation towards economic activity in agriculture for some of the poorest and food insecure populations in the world is the best possible choice. Agroforestry, being a technology, is a factor of partially capital but mostly innovation. And productivity is the only sustainable means of elevating per capita income and thus living standards sustainably long term. Given that agriculture is the dominant employing sector in Africa, a boost in productivity that is at the same time only positive as far as externalities for the environment and local biodiversity is a clear win that can significantly increase well being. Also, it can be noted that experimental economics has revealed that a moderate direct injection of money directly to these populations takes up slack in aggregate demand and increases economic activity, i.e., the money supply is highly endogenous, suggesting a solid multiplier.

In addition, applying AI to optimize the use of this technology provides an excellent opportunity: many are concerned of an AI bubble, and Acemoglu’s pessimistic note on productivity gains in AI thus far don’t bode well for those sky high evaluations. One observation is that they often assist the class of Bullshit Jobs described above. However, by funneling this oversupply of technology directly towards maximizing TFP gains from Agroforestry, this, in Schumpeter’s model, could constitute sufficient disruptive innovation to potentially set off a highly expansive creative Kondratieff economic macro-cycle. And so this sort of research can potentially have the greatest possible positive impact on human welfare, in my humble opinion (and if any EA donors are reading, wink wink).

And as this blog will blend into my professional academic research, I intend to report on the most interesting and novel of results in the latter, and try to relate it to some contemporary phenomenon people can readily understand. Practically, this sort of thing can help with citations as well as open the door to inter-disciplinary collaborations, which I am generally increasingly more interested in exploring.

Finally, I’ve lived a full life, having traveled to about 30 countries, living for at least 6 months in 5, having a period of time involved with various exciting outdoor sports such as rock climbing, spelunking, and mountaineering and others, am a multi-instrumentalist musician, and also have had a wild personal past. As such, far from ivory tower isolation, I have had a wide range of life experiences to draw from, interactions with people from various walks of life, experiences at the edges of human sensory input, and wisdom from many lessons in failures and mistakes over the years, as well as triumphs over struggles.

But this beginning follows a recent victory, one of accomplishing a dream I had for almost twenty years. I am now tenured, that is, on a permanent contract, at a university as a Researcher, that is, without teaching duties. An added benefit is the sense of freedom of speech that I can count on in these troubled times. Classically, the intention of tenure is to permit a researcher who has demonstrated technical proficiency in the field to pursue high risk high reward research, that is, exactly the type that leads to prosocial innovation. So in this sense I hope this blog assists to fulfill my civic duty. Certainly beyond writing I hope to learn, from comments and discussions as well as in the process of formulating the articles, and it shall help my research, in particular of the inter-disciplinary kind..

Upcoming Articles

The next article will describe my own personal biases/priors, as far as preferred Philosophers and philosophical frameworks, beliefs regarding the most optimal conduct of research, as well as my own ideological beliefs as far as politics, values, and spirituality. This transparency, I believe, is a necessary step in epistemic virtue. I also understand that my foundations are unique and unusual, and I welcome readers to consider that ultimately, the map is never the territory, and, since much of intellectual discourse is hermeneutic (what this sentence means will be a topic for an article), what I write about may not be for you.

Afterwards I will proceed with regular article writing on topics of meta-applied-mathematics, meta-science, science and ideology, history, philosophy, economics, and even sociology and anthropology. Although I will describe the two tragedies I encountered that led me to all this in two separate articles, that is only in service of the generalist intellectual discourse. As much as I thoroughly dislike the individual, I do not recommend anyone take offense to him in particular, rather, as a canary in the coalmine, the experience led me to significant insight that I can be grateful for. The insight provided exhilaration, in Sartre’s presentation of a visceral sense of life finally actually happening. As a result of these events, a mission and a vision appeared of a better world that, in part, can be served radically frank but properly in depth intellectual discourse in the liminal space of the inter-disciplinary.

---