This is a corollary of be the best; replace SIGGRAPH with CHI/UIST for HCI projects, ICCV/CVPR for vision projects, etc.
I usually propose safer/clearer ideas for collaboration, especially with more junior folks, and reserve riskier ideas for myself.
Sometimes it is faster for me to directly code than to indirectly describe what to do, especially for vague thoughts. (Some of my single-authored SIGGRAPH papers are results of not finding any collaborators.)
Given these, if a project cannot be published in a top venue, it is usually due to (lack of) execution, e.g., a first-author student needed to graduate and thus ran out of time, or chose to follow his/her own opinion instead of my advice.
I am completely OK with all these; I do the best I can without worrying about factors beyond my control.
Leave a Reply