CRACKPOTS #40: Science at the End of the World (Patreon)

This edition of Crackpots was written by Dr. Marianne Reddan, a cognitive neuroscientist and director of the Socioecological Cognitive Neuroscience Lab. Nothing she ever says or does reflects the opinions or goals of any institutions or employer. Everything stated in this essay is her personal opinion and purely her own. She recently started a science-comedy project to increase both public engagement with science and scientific engagement with the public. It is called Good Brain and it has a monthly live show slot at Caveat in the Lower East Side of New York City. Come check it out and tell her your thoughts.

---------

Life is a fantastic rarity in our universe. I am confident about this, though I will never know if it is Objective Truth. Life does not have an objective definition, nor a subjectively agreed-upon one, and we do not have access to the vastness of the universe in order to calculate its likelihood. Still, we live it and behold it as good.

Science is often framed as a hunt for life’s Objective Truths. But while Truth may be the aim, reaching it in its totality is simply not possible. Science is not the possession of Truth, it is the art of not knowing. Science is a method of inquiry, a set of principles we apply to help us navigate the dark. It is wholly incompatible with dominion, and anyone who gets too cocky with it will be knocked down (á la Jurassic Park). Ultimately, it is a collective endeavor that deepens our existence and helps us make sense out of life. It also put us on the moon, developed the computer you’re reading this on, and cured a whole load of awful diseases. Despite all this, a very deliberate choice is being made to cut this fruitful public service. DOGE claims to be liberating science from the shackles of bias and ideology. Instead they are strategically defunding hard-earned grants based on DOGE’s bias and ideology. But I digress.

I have three reasons for writing this. First, I want to argue that The Quest for Objective Truth is an ideological project that has slowed scientific progress and soured our relationship with nature. The problem being The Quest, not the Truth. Second, I want to report from the ground, in my own vague way, how utterly terrifying it has been to be a scientist in the United States since February 2025. And last, I want to shit-talk human dominion, the AI race, and the glorification of LLMs, as I gently remind you that much deeper intelligence always surrounds us; it just speaks in unfamiliar tongues.

The False Flag of Objectivity

Objectivity exists, but we do not live in the Objective Universe. We live in an interaction among approximations. What do I mean by that? Our primary environment is human society. Human society emerges from the natural world, and therefore its existence depends on the conditions of the natural world. However, as layers emerge out (take for example how software emerges from hardware, and how game strategies emerge from rules coded into the software, etc.) all information necessary to understand the behavior of something within an emergent layer is contained in that emergent layer. You do not need to know how to write the Python code for Minesweeper to win at Minesweeper. Nor do you need to know how to build the GPU of the computer you play it on. Are we interested in understanding the hardware? Absolutely. Does it matter if the task is to win at Minesweeper? No…unless someone spilt water all over your machine and it started acting funny, but that's a digression for a different essay. Human life and society are Minesweeper in this example. The information-space we wade through with our senses is primarily determined by a society of brains making their own approximations of what is in front of them.

Take this podcast you love to listen to—it has no objective value. The acoustic frequencies of the recordings nor the DNA of Brace, Liz, and Yung Chomsky code for how much TrueAnon’s ideas and labor cost, nor how much you are going to like it. Your brain does, and it does it while interacting with many other brains, all of which are operating within social and economic systems. 

Everything living is approximating something. It’s how living things get their needs met. Consider, for example, a chair. My brain will never contain a true one-to-one representation of a chair, but it can approximate it and pull out the gist from all approximations of all instances of chairs it has experienced to form the concept of Chair, a concept that includes how I use and relate to chairs. The Chair is something sturdy I can use to sit and to rest without sleeping. This object has value and meaning to me, as an organism that sometimes must sit and rest without sleeping. Proponents of embodied cognition theory would go so far as to claim these Chair “affordances,” or the actions an object or event begets, are simulated inside the somatosensory and motor planning cortices of my brain whenever I see a chair. Cognitive scientists claim that I cannot have the abstract concept of Chair (or of Science, Love, Cat, etc.) without it being grounded in some type of sensory and motor simulation. Otherwise, what use would it be? How could I understand it? Interact with it? Recreate it? In this sense, the Chair is born out of my relationship to it, and this relationship is personal. Ubiquitous and mundane, but still personal. Still mine.

ChatGPT, or any current LLM, cannot develop abstract knowledge or reasoning because it does not integrate multiple modalities of sensory information (i.e., auditory, visual, touch, etc.) or act on its environment. Taking action in your environment is itself a dialectic. Your action was determined based on some prediction in your brain and the outcome of your action gave you feedback about whether your prediction was correct. LLMs cannot engage with the world like this yet. Instead, LLMs are fed disembodied abstractions (words) and learn the probabilities that connect them. It is sort of cheating…. We earn our words through our stumbling experience of the world. Possessing an abstract symbol without all the simulations attached is not knowledge. It is like cramming for a multiple choice exam rather than mulling over ideas to create a new essay. Cramming is a good strategy if your only goal is to pass the exam. Mulling it over is a good strategy if your goal is to deepen your understanding. What’s worse is that LLMs are often trained on the same data they are tested on, so even their beautiful probability chains are, in some examples, cheating (see Mirzadeh et al., 2024). 

I am often perplexed by the attention LLMs receive from psychology and neuroscience. There seems to be a naïve narcissism underlying their appeal. That is, if they were not language models, if they communicated in some other form, I do not think they would be received as Intelligent™ by their current champions. Chatbots reproduce our language, echo our culture, and mirror ourselves back to us. We think they are smart, because we see ourselves as smart. This type of vanity gets in the way of Truth.

There is a massive amount of intelligence all around us that we callously disregard because it doesn’t serve us or because we are swimming in the ideology of human dominance and exceptionalism. Take, for example, the cat. The cat can plan and execute incredibly sophisticated jumps. It can adjust its vocal frequencies to optimize for human attention. It can respond to your call or ignore it. It can purr on your chest when you’re having a bad day and lay over your hands when your piano lessons have gone on too long. They communicate with you, but they do not use the human language to do it. They use highly evolved social signals, and they can code switch their signals based on who the intended receiver is (e.g., human, cat, or dog). They do all of this without causing catastrophic pollution and energy failures in Memphis, Tennessee; yet no one is afraid of the cat takeover or is seeking to pour billions of dollars into their maintenance and growth. Just the opposite, people are all too quick to deny them sentience while ascribing it to Grok.

This is not surprising. LLMs are, in many ways, a vanity project much like the Byzantine crypt in the basement of the Met Cloisters. Ask someone who works there where the bodies once held in those stone chambers are now. No one knows. They were tossed and the relics sold for display. Ask someone at Meta where the training data came from. Same result.

The chair you are sitting on belongs to you only so long as you are sitting on it. You can keep the representation though.

The Unexceptional Intelligence of Other Beings

Ideology permeates everything—including science. We’ve reached a moment of utterly bizarre doublespeak, wherein science is being gutted as punishment for its recent efforts to erode the longstanding economic barriers to participating in it. This is ironically being lambasted as a non-scientific ideological folly, when for centuries science was steeped so thoroughly in dominator ideologies that we couldn’t even see how the False Virtue of Objectivity was obscuring our view.

The ideology of domination necessitates closing off, rather than opening up to new evidence, ideas, and methods of inquiry. It is therefore in direct opposition to good science, and we have been unknowingly suffering its consequences for centuries. For example, neuroscience was stalled by the Myth of the Homunculus— that an executive power must exist within the brain to control the other parts of the brain. Dominator logic birthed a fallacy of infinite homunculi rather than a useful theory of human brain function and cognition. In ecological networks, competition and predation are no more common than cooperation and mutualism, yet over 67% of publications in the journal Ecology study “competition,” and less than 2% study “cooperation” (Raerinne, 2020).

Recently I met a computational ecologist who builds networks of differential equations to model pollinator/plant relationships. While studying a network of flowers and insects where some insects can consume all flowers and some only consume one type, she found that her model failed to explain real world observations. When all pollinators consume opportunistically, the model predicts ecological collapse because bees that can consume many flowers crowd out the bugs that only consume one type. In reality, however, the network is thriving. After mulling it over she decided to model the bees as adaptive agents; as something that could evaluate current conditions and make a choice. It was not until she allowed the bees in her model to have adaptive choice did she crack the code and uncover a math that reflects back the reality of the ecological network. This is an act of True Science. In order for her to reach this conclusion, an ideological veil had to be lifted. The dominator ideology delayed the discovery of this truth because this ideology does not allow for the possibility of non-human animals to possess cognitive abilities. Being that we are still captives of this ideology, this story does not have a happy ending. Now that bee cognition and the collective intelligence of hives and swarms are acceptable topics of inquiry, this type of research is being exploited for military drone operations. So it goes.

The Terror of Keeping My Mouth Shut

Since February 2025, scientists have been advised not to say anything publicly that could garner negative attention from the administration. Though I kept my head down and my mouth shut, even when it felt cowardly and hypocritical to do so, my already limited grant funds were cut. I was—and am—really worried about losing my job and about losing decades of scientific progress.

Žižek warns us that the misalignment of experience and ideology results in alienation and rage. In that basement crypt of the Met Cloisters, the museum curators placed a small mirror, etched with the words, “Everything will be taken away.” If you pull up for a selfie, looted sarcophagi reflect back at you with those words overlaid. I presume the curators wanted the mirror to remind us of the fragility of our existence, but all I see is the dominator model laughing: The owning class will shake you down to your coffin. 

I met the brilliant and unnamed ecologist referred to above while visiting the Santa Fe Institute. There I spent four weeks in the high desert of New Mexico with fifty-five brilliant complexity scientists, each living and working in the undergraduate dorms of the Institute of American Indian Arts (IAIA) on projects that ranged from optimizing the local fire department’s response times to understanding death as a phase transition to using chaos theory to stop space engines from blowing up. They worked creatively and collaboratively at all hours. They made time to teach what they know, and awarded one another for admitting lack of understanding. Do you know what happens when you can freely admit you are utterly lost as your colleague manically writes equations on a white board and still be appreciated as intelligent and valuable? The veil of the dominator ideology blows off. In the IAIA bookstore, an employee carefully hung a metallic reproduction sticker of the artist Jeffrey Gibson’s “It never did belong to me.” It was for sale. I bought two.

If this is science at the end of the world, we have cause for hope.

Striving for objectivity doesn’t require ideological neutrality; rather, it can be achieved—and even enhanced—by consciously engaging with and balancing competing ideological positions. Science belongs in politics, because science, when loved as a process rather than commanded as a hunt, requires humility, openness, and the trading in of veils for Truth. When science is gutted, there is much more than technological advancement at stake. Our relationship to our existence is weakened. If I never wrestled with the veil, I may never have come to Know that life is fantastic, beautiful, and worth protecting. And it is this scientist’s opinion that if we do not let go of the veil, we may be dragged with it out of existence.