MaxxBench introduces 89 modified DeepSWE tasks as a new benchmark

A post by Ethan Mollick (@emollick) on X introduces MaxxBench, a benchmark of 89 DeepSWE tasks with small changes to the instructions or starting code.

The post's own wording is the only source reported, so the post is the basis for the benchmark's existence and size, and the details of the changes are not described beyond that.