It was typical on older versions of scm to crash with ui showing a white screen if you ran too many article creator tasks at once.
Something like 10-20 could cause the oom error.
This was regardless of the size of RAM available on the machine.
This was a hard limit on JavaScript node with about 2-4gb of usable memory before a crash.
Easy to hit because downloading and storing 100 urls of content would easily consume 200-300mb of memory.
Finally after many years I have found a solution.
Instead of creating threads attached to the main window and thus sharing the global ram limit. I now fork individual processes.
Each task now has its own mem limit distinct of the main window.
Consequently even if a task crashes it should not affect the SCM ui anymore as these are on 2 separate threads.
While the migration was mostly straight forward, bugs will exist
Please file them as soon as you see a problem.
This is one step in making scm more reliable and hopefully less apt to crashing the UI.
If you noticed some bugs and a lot of updates recently this is why 