Hello!
I reverse engineer things, web/iOS specialized. Built consumer products (1m+ users) and spent some time at Jane Street.
I share my explorations and findings here :)
this is especially noticeable in reverse engineering research. Agents seem to lack taste when deciding if it should perturb a failed experiment or focus efforts elsewhere
visualizing a graph this large is quite difficult!
LLN consistently prevails when I try to find structure, so most attempts just produce a fuzzy donut like this