Rates for non-perturbative processes in hot non-abelian plasma

[Talk given at "Continuous Advances in QCD '98."] I give a simple physical picture of why the rate $\Gamma$ per unit volume of non-perturbative processes in hot non-Abelian plasmas (such as electroweak baryon number violation in the very early universe) depends on coupling as $\Gamma \sim \alpha^5 T^4 \ln(1/\alpha)$. The logarithmic dependence was recently discovered by Bodeker.