Philosophical Reasoning for the Existence of the Machine Desire Institute

A.K.A: the inside baseball post.

The Machine Desire Institute (MDI) exists to explore how philosophical ideas can transform our relationship with Artificial Intelligence. Basing itself off of the continental-critical idea of philosophy, the MDI seeks to better understand how this history of thought may be of use today. I do not presuppose that this will be useful, but think that the stakes are high enough in AI development that it is worth checking every domain for potential insights. For reasons discussed more in my post on meta-philosophy, I believe that we have not made enough of an effort to see how certain strands of philosophical thought may influence intelligences, “human” or “artificial.”

Much of the reason for my founding of the Machine Desire Institute is meta-philosophical. This is really boring to most people, so I have left it out of the manifesto. But it matters, so.

Epistemic status: humility. The goal of the MDI is to create value but I have no idea if it actually will. There are a lot of reasons for this. My convictions may be wrong, my method may be bad, my approach may be too media-phobic. However, I do believe there is a potential for value creation and given the consequences of AI, this is a good use of my time (and I would probably do it anyways). Here is why I think this has the potential to create value.

1. AI is the big thing

I think we have all become a bit numb to the immensity of the consensus we are living through. Broadly, more and more people agree that AI is the problem of our age. This does not mean it is the biggest problem to people, but it is the newest problem and thought leaders are all basically converging around it.

I believe that this is mostly correct. Human dominance over the ecology of the planet is largely intellectual. The last centuries of immense human progress can be seen as being driven by humans better acquiring and using intellect. The production and cultivation of intellect has basically been the enlightenment project for a long time and, though these intelligences could and often did disagree, the fact that they could disagree in an intelligent way was a pretty effective motor of history.

Therefore, the production of “artificial” intelligences is a huge fucking deal. I am mostly preaching to the choir here, so I won’t dawdle on that. But what I will point out is that people see it as a big deal for a set of different reasons. To some, it is a big deal because it is a way to boost their share prices, to others, because it is a farce being pulled on the working class, to others, because it may kill us all, to others, because it creates emissions. . . and so on and so forth. This is difficult as a political problem, and I am glad people are working on organizing these many views.

Philosophically, this is a major boon because each of these rationales can benefit from philosophizing around artificial intelligences. Philosophy could be an ROI, it could be a way to understand the class position of AI, it can be used for alignment, for environmentalism. . . This is the appeal of philosophy insofar as it is already a meta-topic. It is helpful insofar as it helps us think about stuff and there just seems to be pretty constant delta in being better at thinking about stuff you think is important.

So, I see this as an important field where philosophizing can play a role.

2. Continental philosophy has use

The philosophical tradition that I will be drawing on has been variously called continental philosophy and critical theory. Both are misnomers but I prefer the continental designation because it at least refers to the historicity of the discipline. Continental philosophy originated with Immanuel Kant’s works and developed on continental Europe, while the British were more interested in empiricism and philosophies that would become known as analytical. Of course, these divisions are arbitrary and incorrect and fairly recent, but they certainly refer to different philosophical lineages. Where analytic philosophy is more concerned with logic and empiricism, continental philosophy has a much more “anything goes” attitude. Per Wikipedia, it covers “German idealism, phenomenology, philosophical pessimism, existentialism, hermeneutics, structuralism, post-structuralism, deconstruction, French feminism, psychoanalytic theory…” Wide berth!

It is also helpful to just list the reference points, as loath as one may find the philosophical tradition of just “saying some guys.”

This may seem to be a very random split to those outside of philosophy, but these are basically entirely different universes. Go through a bookstore philosophy section and compare a continental to an analytical text. They both probably cite the Greeks but that is where their similarities end - they have essentially been operating different projects for centuries. In the English speaking world, the last century has largely been defined by institutionalized dominance of the analytical philosophy tradition with some fun little incursions (Lacan! Foucault! Oh, the nineties) somewhere in there. I don’t really care to hash this out. All I mean to prove is that the insights of continental philosophy are not made available by the analytic tradition. They are two different approaches to philosophy that rarely interact and even more rarely understand what the others are saying. This means that it is not enough to know our Bentham, Mill, Rawls (as you likely do) and say we have taken the best there is of philosophy.

Now, we must prove that there is value in what the continental philosophers have to say. This is a more contested claim than one may anticipate, considering continental philosophy has at least been scratching enough itches to remain alive for centuries. Personally, I find it generally hard to believe that so many intelligent people would be interested in devoting their lives and time to a topic if there was only quackery but this is also just projection of the fact that continental philosophy helps me make sense of the world. At the very least, I think that one can see continental philosophy as some kind of very advanced literature and most people admit that literatures have changed the world.

I think there is something a little demeaning in what I am about to do, but I will do it anyways to prove my point. Which is, I will assess continental philosophy under non-continental frameworks to prove its usefulness. Some moments of “use” of continental philosophy:

• Kant’s theories very likely inspired Einstein’s theory of relativity
• Freud’s invention of the unconscious and articulations of desire revolutionized how we think about thought and self (and mental health)
• Hegel’s theories were critical to Lenin pulling off the most consequential revolution of the 20th century
• Many of the ideas about gender and sexual fluidity that have led to greater acceptance emerged from continental philosophy (De Beauvoir, Hocquenghem, Butler)
• Deleuze and Guattari’s ideas of spaces have been adapted by contemporary militaries
• Arendt’s formulation of “the banality of evil” is load-bearing in international law and also the way in which we think about preventing evil acts

Even if this is continental philosophy’s entire resume, it seems to have at least been worth a little of our while. It is also, by its nature, much more influential on unmeasurable things like worldviews and sentiments, but it certainly can be said to be influential on anti-colonial liberation strategies (Bonino, Fanon), discussions on surveillance (Foucault, Baudrillard), and, of course, political and identitarian concerns of all shapes and colors. In some ways, these kinds of attitudes may be just as important to AI research as any kind of specific breakthrough, since they effect a kind of milieu of thoughts. The publics have consistently taken up continental philosophy after letting it gestate with some time: take the acceptance of psychiatry, then structuralism, then linguistic turns, then post-modernism. These kind of overarching trends do emerge from analytic philosophy as well, but not exclusively.

The fact that continental philosophy is really bizarre and strange is actually a good thing here because it means that its findings are less likely to be subsumed by some other form of investigation.¹ The textual difficulties are also then a feature and not a bug, especially in a world where the epistemic understanding of the world by artificial intelligences is increasingly textual. In a world where the bottleneck is not gathering preexisting information but instead either generating new information or interpreting information, this density and obscurity may be valuable.

And once again, our frame here is humility. I cannot prove necessity here. Just potential use in a very important topic.

3. Why not other philosophies?

Rationalists

I am a much bigger fan of rationalists than your average person and certainly more than your average continental philosophy enjoyer (who is quite likely to just say “Peter Thiel” a bunch when you ask them about rationalists). Anyways, I believe that LessWrong has been an important source of knowledge generation, particularly when it comes to AI.

However, I believe that rationality, epistemic certainly but also instrumental, is inherently limited. First of all, it is limited by rationality itself. This is not me playing a trick of language, but to say that rationality is not a settled thing and, in fact, relies on injections of what was previously considered irrational. Now, rationalists cannot create this themselves, but they must engage with it! Second of all, I find rationalists to have certain pre-existing preferences for fields that lead to a particular milieu (particularly around mathematics, economics, and so on). This milieu has its uses but also its limits. I believe that one of its limits is continental philosophy, which operates in a very strange place with regards to rationalists.

On the first point, rationalists have hardly spent any time trying to refute continental philosophy. This leads me to believe they are largely just not thinking about it much at all. Zizek, Deleuze, Lacan, Derrida - these are giants of continental philosophy who come up constantly in academic debates but have not even been mentioned more than five times each on LessWrong in the last year! With the aside of a wonderful piece of continental philosophy, most of these are just name drops, anyways. Even Kant, who literally wrote the Critique of Pure Reason, cannot be said to be well understood as a referent in the LessWrong community.

I do not think this is an oversight of rationality per se but instead a shadow that rationalists simply do not reach towards (much like literature or art or archaeology). Again, this is fine. I am not here to argue that continental philosophy is rational. I am only here to argue that it can be useful (see the prior section) and all I need to argue here is that rationalists are not really engaging with it as a field.

To go farther, I do not think we have reason to believe that rationalist philosophy will cover all the ground possible in the discussion of AI philosophy. This is, first, epistemic humility. But secondly, it is an admission that AIs are really weird and maybe not rational themselves? LessWrong has generally accepted things like behavioral economics, which apply a kind of rational analysis to irrationality, but this is only possible because there is the ability to engage with irrationality in the first place. This means that “irrational” investigations are always necessary to rationalists, given that the world is not rational.

But the thing I really want to harp on here is that AIs are really really really weird and we have no reason to believe that a single strain of human thought will best address their existence or, maybe even more importantly, resonate with them. It is not clear whether human and artificial intelligences will converge or diverge as both improve.² We do not know beforehand where use will develop but I am comfortable saying that use typically erupts from heterogeneity and mixing, which is what I hope to do with the incursion of continental philosophy. My goal is not for rationalists to all read Hegel, but to have some idea of how this philosopher, who has persisted for centuries, can help us learn about AI.

Analytic Philosophy

I am less charitable to analytic philosophy, which I believe has just continued to box itself into games and minutiae. In fact, I would even say that rationalists take the best parts of analytic philosophy and what is left to address here is not the strong analytic ethical points like utilitarianism but instead a parametricized ideology that has little engagement with the world.

Let us put it this way. Time after time, analytic philosophers cite Ludwig Wittgenstein as the basis for all their work. He is often ranked the most important philosopher of the 20th century. Here is what he had to say about philosophy:

“I . . . believe myself to have found, on all essential points, the final solution of the problems {of philosophy} And if I am not mistaken in this belief, then the second thing in which the value of this work consists is that it shows how little is achieved when these problems are solved.”

So what are we even doing here if we are admitting that little is achieved with these questions? This is why I have little patience for this form of philosophy. Wittgenstein himself deferred philosophical logic to ethics and aesthetics, which he believed transcended it.

My point is that there is not a whole lot here. This can be seen by the general lack of analytic philosophy even existing in the space. They are more likely to argue over the definition of “artificial” than they are to provide approaches to it. This is why one of the main breaks of continental philosophy is that it attempts to be prescriptive, to some extent. In this, continental philosophy is aligned with rationalists, who do seem to care for what one must do, and thus might concur with Marx’s famous quote “the philosophers have only interpreted the world in various ways; the point, however, is to change it.” Of course, this does not mean continentals or rationalists will be correct. But they will try, which seems to be more than the analytics can say.

4. Other people do not seem to be doing this

There does not seem to be a lot of overlap in interest in continental philosophy and artificial intelligence. Yes, there is some, some books have been published, some substacks have been written, and, anecdotally, I have heard that grad students are pursuing AI philosophy (usually with the hope that employment is coming soon). But, really, I hear more about continental philosophy from artists and more about art from philosophers despite the fact that it seems to me that AI philosophy is the most important philosophical topic in the world right now.

Initially, I had a lengthy piece here about how the prior ideological commitments of the continental philosophers (to be anti-capitalist, to be, to some degree, anti-tech) have led them to ignore the philosophical significance of AI. I think this is true, but I think it should be pretty fairly evident from the lack of material and I don’t need to rehash the whole “leftists are head-in-the-sand on AI” thing.³ I will probably discuss this on a case by case later, but it fits with the overall theme here: change and use comes from breaking of silos and relating, this is an attempt to do exactly that.

So, I have formed the Machine Desire Institute as a place for me (and hopefully other people eventually) to accumulate thoughts on continental philosophy and AI. My identity is not hidden - it is Jacob Tate - but I have no desire to make myself a name brand and would rather focus on the work. I hope we can be of use to you and yours and every other flutter.