Struggling to find anything interesting or non-obvious about this article. You give a bunch of LLMs various parallelizable task and some models manage to do it well but others don't. No insights as to why. As someone with a distributed systems background the supposed 'insights' from distributed computing are almost trivial.