It has been proposed that inertia can be explained as follows: when objects accelerate in one direction a Rindler horizon forms in the other direction, suppressing Unruh radiation on that side, and producing a net Unruh radiation pressure that always opposes the acceleration, just like inertia. So far this model has predicted masses over twice those expected. In this paper an error in this model is corrected so that its prediction improves to within 29\% of the expected Planck mass. It is also shown that inertial mass may be understood qualitatively by applying Carnot's principle and entropy to Unruh temperatures, so that the work needed for inertia comes from the difference in the Unruh temperatures seen by the accelerated object and the cosmos. This implies that highly-accelerated systems may emit heat in a new way.