Notes

I started my PhD in one technological epoch and finished it in another — why the bottleneck in machine learning for science is measurement rather than hypotheses, and the people who lit the path. Adapted from the introduction and acknowledgments of my thesis.