I got a temp job in an official statistics department at the age of 23. Somehow; while I was nominally educated (economics & philosophy), I'd never written any code, never simulated anything, never calculated a PDF, and never actually analysed more than 50 rows of data.
Most notably, I'd never heard of principal components analysis (nor any unsupervised learning, tools with which one can discover the 'joints' of nature without really putting one's mental framework into the mix).
We (they) built a PCA model of childhood poverty, letting the algorithm find the most important parts of the phenomenon, making them fall right out of the data, heedless of the Marxism or Malthusianism or indeed mindlessness of the modellers. Despite my tiresome anti-realism and anti-quantoid convictions, I could see the philosophical implications were huge. So here I am.