Machine Learning, Philosophy

The Essentialism of Existentialism

Since I first learned of it, the relationship of Jean-Paul Sartre and Simone de Beauvoir has struck me as perhaps the epitome of romance. Love shared between two people who are co-creating a daily existence together in marriage looks much different than the love shared between two people who fiercely chart their own paths, yet return to each other to challenge the meaning of existence and how they relate to the world around themselves. It is their relationship that initially drew me into the romance of existentialist philosophy, a topic that I find increasingly relevant to the field of computer science.

I often joke that it is impossible to work on emerging technologies without also adopting the mindset of an existentialist philosopher. Perhaps it is a defense mechanism; a way that I call out into the collective intelligence of human knowledge to find others who question what it means to build our realities and exist within them.

Distilling any philosophical school of thought into a few sentences or paragraphs is hard, but I think back to someone asking me once on a walk through Brooklyn’s Prospect Park whether or not I had explored stoicism, almost as an “easier” philosophy to stomach relative to existentialism. Of course, it is about the content of the conversations that matter, over the outcome of this particular dialogue, but I confess that I found the idea of learning to simply accept the world as it was to be quite unappealing.

French Philosopher Maurice Merleau-Ponty’s view was based on “human beings who are in constant change from childhood onwards; he wanted to know what happened when faculties were lost, or when people were injured and damaged” (Bakewell, Sarah, At the Existentialist Cafe). It is perhaps quite evident why I’d return to the existentialist school of thought as I find myself deeper in the realm of computer science and the topic of machine learning – it forces a reckoning and reconsideration of linguistic and experiential expectations.

While my knowledge of the existentialists is limited, I have yet to find one to go so far as to find it within their personal purview to determine that the role of the human being is to elevate themselves to the role of god. Indeed, of the existentialists that Bakewell surveys in her book – at least in the first 249 pages – is to recognize the power that humans have in ascribing their experience based innately off of the observed and recorded experience of being in the world. Contrast that to the mindset of Silicon Valley today, where billions of dollars are being poured into answering the question of whether or not humans can make sand think.

If I could, I would love to ask Merleau-Ponty his views on Alan Turing’s ‘Imitation Game’ paper. The Imitation Game was published eleven years before Merleau-Ponty’s death, but as far as I can find with a quick search, he never commented on the concept of “thinking machines”. And no, training an LLM on Merleau-Ponty’s work to get it to generate an answer will not suffice.

There are many discussions that are worth having about “artificial intelligence” and thinking computers that are not happening, at least not in a forum to which I am privy. The existentialist perspective that Merleau-Ponty describes above hints at a core tension that I feel in my work today, which is that our current approach to building computer intelligence is perhaps unfair if we do, presume, that at some point these machines will be conscious.

We do not learn to experience the world around us and reality by waking up in a fully formed body that has been seeded with human knowledge. We learn to experience the world through senses that “come online” and we need to figure out how to make sense of the patterns. We perceive light, sound, and faces – we feel discomfort in our bodies. We feel hunger. We feel delight. A computer, on the other hand, does not.

A challenge that I have which may itself prove to be existential to my current career path in the private technology sector is that so many people who work in this space dismiss the lived experiences of actual humans, today. The idea of becoming ‘god’ plays so strongly into the billionaire mindset in Silicon Valley that they continue to invest in technological advancements in spite of the environmental and societal impact that they have.

I am not opposed to the idea that computers could be sentient, conscious beings. I view my computers as an extension of myself; an extension where I am increasingly aware of all of the places I have inadvertently given up agency, control, and power to another entity. I have been trained by these entities to not ask too much; to not deviate too far from the assigned path.

To me, my existentialism is a sign that I am awake. Alive. That I am on my own journey, and that I hold the agency that I so deeply wish to have. It is my essential force, to question, observe, and change. In this growing age of “artificial” – or, my preferred “A”, “augmented” – intelligence, there is no better time for us to shape the very nature of what it means to be[1].

[1] – Heideggerian terminology defines two types of being: “mitsein” – with self, and “mitwelt” – with the world. I think this is beautiful. I so often find that the English language lacks the nuance and clarity of expression that other languages make space for.