Why do ice sheets break into such big pieces? Another massive chunk of the Antarctic ice sheet broke free. Giant crack frees a massive iceberg in Antarctica. The article has a map that shows more cracks. It looks reasonable that they will lead to future breakup.
What sets the scale of these pieces? Why so large?
This post, Does the ice melt faster when breaks free from larger sheet and drops into Antarctic Ocean waters?, references this NASA paper, Warm Ocean, Not Icebergs, Causing Most of Antarctic Ice Shelves' Mass Loss. It says that basal melting causes more loss than calving. I understand that calving from glaciers is the source of smaller icebergs.
The large pieces seem to be something different. A flat floating ice sheet develops cracks.
Do the cracks come from stresses due to ocean currents? Thermal changes? I am sure thinning ice must have something to do with it.
Do the cracks follow lines of weakness, something like earthquake faults? What would be the source of such a weakness? Or perhaps it is more like a crack propagating in breaking glass?
 A: I've done some digging and this turns out to be a really interesting question.  In this answer I am basically reporting on what I have found.
In a 2003 paper, it is stated that mechanisms for large iceberg formation is still not understood well:

Mechanisms have been proposed to explain the calving of small icebergs (with width about equal to one ice
thickness) from the ice fronts of tidewater and lake-bounded glaciers (Reeh, 1968; Fastook and Schmidt,
1982). However, no coherent theory exists which is concerned with the calving of large tabular icebergs
from Antarctic ice shelves (with main axis lengths of tens to hundreds of kilometres). Such calving events
appear to follow a certain basic pattern: inlets open more or less perpendicular to the ice shelf front,
intermittently widening and growing, until the final break-off of icebergs occurs rapidly, and frequently, at
an angle with the original orientation of the inlet (sometimes joining two inlets).

But now, I've come across a couple of papers that have proposed at least one surprising mechanism: far away tropical storms.
See this 2006 paper and the following quote from section 4:

Considering the long history of investigation into trans‐oceanic propagation of long‐period ocean waves [Munk et al., 1963; Snodgrass et al., 1966], it is not surprising that our observations have revealed examples of sea swell traveling half‐way around the earth to shake icebergs and ice shelves along a broad swath of the Antarctic coastline. What is unique about the observations presented here is that they imply that giant icebergs and the calving margins of major ice shelves are mechanically influenced by meteorological conditions in the far field. Considering that iceberg calving is involved in the maintenance of Antarctica's ice‐mass budget [Jacobs et al., 1992], it is thus natural to consider whether climate conditions in the extra‐tropical northern hemisphere and the tropics (e.g., storm intensity or hurricane/typhoon frequency) could exert a control on the mass budget of the Antarctic Ice Sheet.
[9] A leading hypothesis for how tabular icebergs are calved
concerns swell‐induced vibration that can fatigue and fracture ice at
weak spots, producing rifts that become iceberg‐detachment
boundaries [Holdsworth and Glynn, 1978; Kristensen et al., 1982;
Zwally et al., 2002]. With this hypothesis in mind, we considered
whether the arrival of swell from the Gulf of Alaska storm discussed
above contributed to the spectacular break‐up of B15A on October 27
2005 (Figure 1a).

So vibrational energy from traveling sea waves is thought to play a major role in the formation of large icebergs in Antarctica.
In 2018 a paper was published in which a researcher attempted to model this using data collected from seismometers.  It was inconclusive.

To the best of my knowledge, no previous study has definitively demonstrated that ocean waves may trigger ice shelf rift propagation. To address this situation, I have attempted in this paper to construct a simple model of wave‐induced rifting. Although I have been able to make this model behave in a manner consistent with observed rift behavior, no large rift propagation event occurred during the period from which I have data. As a result, definitive proof of ocean wave triggering remains elusive.

So the bottom line is: climate scientists don't know yet.  It is an active research question!
