Accounting for the way the Antarctic ice sheet interacts with the solid earth below - an important but previously poorly captured phenomena - reveals that ice sheet collapse events may be delayed for several decades, at this major ice structure. Specifically, this interaction at the so-called "grounding line" may slow the otherwise fast-moving retreat of the Thwaites Glacier, the largest glacier in this region. "For those concerned about potentially catastrophic sea level rise, the results ... can be taken as welcome news," says Eric J. Steig, in a related Perspective. "But it is important to recognize that [the authors] do not make a specific prediction about the magnitude of the West Antarctic contribution to sea level." The question of how quickly the Antarctic ice sheet will lose mass over the next few centuries - contributing to global sea level rise as it does - continues to motivate glaciological research in Antarctica. Recent work on the West Antarctica Ice Sheet (WAIS) has shown that an unexpectedly rapid rebound of areas of this sheet may help stabilize the WAIS against catastrophic collapse. This is based on the way in which, as glaciers retreat, the Earth's crust springs up, creating changes in the sea bed geometry in relation to the point where the glacier goes afloat to become an ice shelf (the grounding line). This process can prevent the steady retreat of the glacier, allowing it to ground. To date, numerical model simulations of this process have been used to assess it at relatively large spatial scales. Here, Eric Larour and colleagues focus on the feedbacks between glacier retreat and solid earth processes at finer resolution, using the Thwaites and Pine Island Glacier grounding lines as case studies. In a sensitivity experiment using model simulations, they find that by the year 2350, the grounding line retreat from the Thwaites Glacier is reduced by about 40%, compared to a scenario where fine-scale feedbacks aren't included. Too, the Thwaites Glacier's contribution to sea level rise is reduced by more than 25%. "...their results should be seen as a guide to the magnitude (and sign) of uncertainty in existing predictions, and as a roadmap for future research," says Steig.