I will present a uniform analysis of the atmospheric escape rate of the known sub-Neptune planets with estimated radius and mass (Mp < 30 M_earth). For each planet we compute the restricted Jeans escape parameter, Lambda (the escape parameter for a hydrogen atom evaluated at the planetary mass, radius, and equilibrium temperature). Lambda is an indicator of extremely high thermally driven mass-loss rate (Lambda < 20). We identify 28 planets (out of 167) that are consistent with H2/He-dominated atmospheres and are expected to exhibit extreme mass-loss rates. This constitutes a contradiction, since the H2/He envelopes cannot be retained given the high mass-loss rates. We further estimate the planetary mass-loss rates (Lhy) with tailored atmospheric hydrodynamic models and compare to the energy-limited high-energy (XUV-driven) mass-loss rates. We confirm that 25 planets (15% of the sample) exhibit high thermally driven mass-loss rates (Lhy > 0.1 M_earth/Gyr), in excess of the XUV-driven mass-loss rates. We hypothesize that either the masses (from transit-timing variations) are being underestimated or that the radii (from optical transit observations) are being overestimated because of high-altitude clouds/hazes. The cloudy-atmosphere scenario is supported by the higher albedos found for super Earths and similar correlations between log(g)--Teq and cloudy/clear atmospheres for hot Jupiters. I will discuss the implications on the overall properties of sub-Neptune planets, particularly on the observed mass--radius distribution.