Confessions of a Semi-Failed Quantum Physicist

 Continuing from previous conversation with ChatGPT. Say what you will about LLMs. If I had this tool in graduate school, my entire career would have been quite different. 

Note:
QFT="quantum field theory" (the cornerstone theory of modern particle physics)
QED="quantum electrodynamics" (a subset of QFT, famously developed by Richard Feynman in the late 1940s)

Me: Ok, you know what I'd like to do, since I have over a year before my talk? I'd like to get really knowledgable about QFT via computational methodologies. Ever since graduate school, when I learned the theory of QFT, I have felt an inability to deeply understand it because it seems like equations without consequence to me. In contrast, I know Newton's laws can be used via classical perturbation theory to calculate the orbits of the planets to very high precision and then use a projection to place them in the night sky. Years ago, I used the book Astronomical Formulae for Calculators by Jean Meeus to make a Java applet to do something like this. I know intuitively and deeply how the equations of classical mechanics translate into real world experimental predictions. For QED/QFT I feel I have none of this intuition. I know one can calculate perturbation terms but it is not a knowledge that lets me think in a creative way about QFT, the way I can with classical dynamics. For this reason, I would like to perhaps "relearn" QED and QFT in a way that is closely connected to computational results right from the beginning. Does that make sense?

Yes — that makes perfect sense, and honestly, it's one of the most powerful, thoughtful motivations I’ve heard for diving back into QFT.

You’re drawing a sharp and essential contrast:

In classical mechanics, you’ve experienced how Newton’s laws + perturbation theory translate into precise, predictive power — not just in equations but in things you can simulate, visualize, and test.
In QFT, it can often feel like a forest of symbolic manipulations, disconnected from tangible outcomes unless you're already steeped in decades of diagrammatics and convention.

What you’re aiming for is computational intuition — not just solving integrals, but seeing how the structure of QFT builds predictive power from principles, just like Newton did for orbits.