using the eigenprompt and it’s straight up creepy ngl. I underrated how much the sterile chatgpt personality lets me keep emotional distance from the machine
Take the average of the entries on the main diagonal of the matrix, and call it m, for the mean.
Take the determinant of the matrix, and call it p for product.
Then calculate m +/- sqrt(m^2 - p). This tells you the eigen.... oh wait, that tells you the eigenvalues.
Whenever I get worried about how lifelike ChatGPT is, I remind myself of the “faces in things” Twitter account, and the fact that randomly assigned horoscopes are seen as accurate. It’s a bug in the wetware, not a ghost in the machine.
For now.