The P2P Reading List (2023 June)
Read the best books first, otherwise you will find you do not have time.—Henry David Thoreau
Theaetetus by Plato
As an aspiring scientist, I am fascinated with inquiries into what knowledge means. In Theaetetus, Plato conjures the soul of Socrates to lead one of his trademarked dialogues about assessing knowledge. Socrates’ conclusion brings no revelation:
"Therefore, Theaetetus, knowledge can be neither perception, nor true belief, nor true belief with the addition of a rational account."
This is indeed disturbing to anyone primed by the modern education system. We expect answers and expect them now. It is simultaneously baffling and humbling to consider that although we could send people onto the Moon and develop a vaccine against covid at lightning speed, it remains a formidable task to define what knowledge means.
The Wise Man's Fear by Patrick Rothfuss
After reading The Name of The Wind last month, I needed more from the Universe of Kvothe—even despite the fact that the third part of the trilogy is nowhere near the horizon.
It is somewhat surprising that fiction inspires deeper thoughts, but this is what the story of Kvothe did for me: I realized I find it extremely alluring that magic in this universe is centered around knowing the true names of things. However, this concept is not out of the blue: as I got to know (from one of this month's books: Getting Past No), in ancient mythology, you could ward off evil spirits by calling their names.
Why does a book on negotiation refer to ancient mythology? Since naming the tactics of your opponent can help defuse it easier. In the social sciences, there is also evidence that labeling your emotions can help handle them better. Names are powerful, indeed.
Noise: A Flaw in Human Judgment by Daniel Kahneman et al.
Nobel laureate Daniel Kahnemann focuses on noise in humans (after the seminal work on cognitive biases, resulting in the must-read Thinking, Fast and Slow) and compares them to algorithms/artificial intelligence. Machines are often better, sometimes even the most simple, horribile dictu, even random models:
“...people are inferior to statistical models in many ways. One of their critical weaknesses is that they are noisy.”
This month's theme might be humility, then. We need to be more aware of our noisy decisions—this is not a case to distrust human decision-makers always and everywhere but to ask whether and when (and in what combination) we can use humans and machines together to eliminate both noise and statistical bias.
With the rise of Large Language Models (LLMs) such as ChatGPT, researchers asked whether these models suffer from the same cognitive biases. In this paper, the authors pinpoint that many cognitive biases described in Thinking, Fast and Slow are present in LLMs, too (the authors used code generation as an example). This is a sobering view, especially given the above (i.e., that sometimes even random models are better than humans, though that statement is mostly about noise). Thus, it is an interesting research (and policy) question whether we want these large models to act like humans (including our flaws) or to be "better", even when that might produce decisions that a human would not make.
I am by no means an expert in the field of LLMs, though I thought about how machine learning models can be used for human decision-making. One point I made is that the users should be made aware of the nature of these predictions (in short: even if the model gives you a single prediction, you should not automatically treat it as something with 100% certainty); for this, it would also help when these systems would be designed in a way that makes this point.
How to skate 10k by Nils van der Pol
This short piece is written by Swedish speed skater, World and Olympic record holder Nils van der Pol and describes his training regimen leading up to the Olympics. As a sports science nerd, I always enjoy reading about different approaches. Though this piece is much more than a dry description of Nils' training regimen: it also describes his ars poetica, with thoughts bordering philosophical depth about the lifestyle of a professional athlete. If you do not care about sports, those few pages I would still recommend—we all can learn from the author to self-reflect better.
The Warrior Ethos by Steven Pressfield
Steven Pressfield, again. I think about this short book as being on the border between the fiction and non-fiction works of the author: it is deeply philosophical about what being a true warrior meant and means in cultures throughout history, but it also connects to Steven Pressfield's fiction books, which frequently (i.e., in all the fiction book I read from him until now) considers warriors, let them be Spartans at Thermopylae, Alexander's Macedonians, or Rommels's Afrika Korps in World War II.
Getting Past No by William Ury
William Ury, together with Roger Fisher, are the greats of negotiation tactics. Getting Past No picks up where Getting to Yes finishes, i.e., when an agreement seems out of reach. The book is short and full of historical examples, which makes reading much more fun. The advice is actionable with simple tricks and tactics, such as calling them by their names.
If you are interested in all the books that piqued my interest, you can also visit my Goodreads profile.