Love the commitment to learning in public — Day 3 of 100 is where most people quietly quit, so massive respect for keeping the momentum going. My advice to my younger self would be: don't just learn the theory, build something small with each concept. Even a tiny project that uses perceptrons to classify something simple will cement the understanding way faster than reading papers. The jump from classical ML to deep learning is where things get really exciting — CNNs and transformers will blow your mind once you have these foundations solid. Keep posting updates!