Yannic Kilcher

Yannic Kilcher publishes long-form video reviews of machine learning papers, often working line-by-line through models, losses, and experimental setups. His commentary blends implementation instincts with skepticism about overclaims, which makes the channel valuable when a paper’s marketing outruns its evidence.

While the runtime is longer than typical “explainer” clips, the payoff is a habit of reading methods carefully: checking baselines, questioning ablations, and spotting ambiguous training details that affect reproducibility. Many practitioners use Kilcher’s videos as guided study when preparing reimplementations or literature reviews.

Follow Kilcher if you want to learn how experienced researchers interrogate new work. The tone assumes audience patience and comfort with notation, so it pairs well with graduate-level coursework or serious self-study rather than casual background viewing alone.