by S. Abbas Raza and ChatGPT 5.1
S. Abbas Raza: You may have heard of John Brockman who is a well-known American literary agent and author specializing in scientific literature. Brockman led a scientific salon for 20 years, asking an annual question to a host of scientists, philosophers, and other thinkers, and publishing their answers in book form. For the last such book, instead of asking for answers to a question, he requested that the contributors each submit one question to which they would like to know the answer, and the resulting list of questions was once again published as a book in 2018 entitled The Last Unknowns with a foreword by Daniel Kahneman. Here is the question that I submitted and was published in that book:
Why is sleep so necessary?

The questions were simply printed in the book with the author’s name, one to a page, without any commentary or explanation but let me give a short explicatory comment here. Much of my curiosity about sleep had to do with the word “so” in my question and the work it is doing there. One reads many accounts of what happens when we sleep. For example, the flushing of toxins from the brain, cellular repair, etc., but what I found unsatisfying was that none of these seemed to me to give good reasons we necessarily have to be unconscious for the process to take place. And the advantage one would gain from not having to be asleep for a very significant portion of one’s life would clearly be so profound that if there were any easy way evolution could have found a way around sleep, over hundreds of millions of years of reproducing different kinds of brains, it would have. But we see that any animal with a brain, no matter how tiny or complex, from fruit flies to humans, needs to sleep. That is why I asked why sleep is “so necessary”, because I suspected that the answer to that question would tell us something very fundamental about the design of the brain.
A couple of days ago, I mentioned to my wife, Margit, that a few months after GPT3 was released to the public by OpenAI at the end of November in 2022, months that I spent intensively learning about how transformer models like ChatGPT work (transformer is the “T” in GPT and a particular architecture of artificial neural networks), it suddenly occurred to me that AI models may have provided us with the conceptual vocabulary to more easily understand what is happening in the brain when we sleep. I then went on to explain to her what I meant in the following way:
Human brains are not “pre-trained” with data the way a large language model (an LLM like ChatGPT) or other types of deep-learning AIs are. Instead, the synaptic weights throughout our “thinking brain”, probably the neocortex at least, are adjusted according to our experiences throughout our life. But while we know what an artificial neural network’s learning algorithm is during pre-training (it’s called “backpropagation“), we know that in our case it is not backpropagation, but we also do not know exactly what the learning algorithm (which goes and adjusts the synaptic weights) for human brains is. Read more »


Preparing a worksheet with negative-number calculations where all the digits are sixes and sevens. Telling myself it’s meant to take the fun out of it for them – like a sex ed teacher having their students say ‘penis’ one hundred times before starting the unit. Definitely not the whole story, but plausible: as a middle school math teacher I am more than justified in trying to tame the phenomenon. In fact, I have drawn a firm line; just seeing a 6 anywhere in an exercise is decidedly not an appropriate reason for doing the meme. Really, we need to get on with the lesson now; I will count to five.
Antonio Gramsci’s theory of hegemony helps explain how the power structure of modern liberal-democratic societies maintains authority without relying on overt force. Many definitions of hegemony point out that it creates “common sense,” the assumptions a society accepts as natural and right.




Art is dangerous. It’s time people remembered that and recognized the fullness of it. For if art is to remain important or even relevant in the current moment, then it’s long past time artists stopped flashing dull claws and pretending they had what it takes to slice through ignorance. We need them swallow their feel-good clichés and to begin sharpening their blades. We need dangerous art, and we cannot afford much more art that its creators believe is dangerous when it is not.
Emma Wilkins’ excellent piece “






