Big and famous groups/companies usually do release their code nowadays. I'm automatically suspicious when someone doesn't, even though I guess most of the time

Big and famous groups/companies usually do release their code nowadays. I'm automatically suspicious when someone doesn't, even though I guess most of the time it's nothing nefarious. It's just extra work and effort to bring the code into presentable shape and that effort could be spent on the next paper. Once the paper is published, the material benefits have been reaped, publication count incremented. Of course this is a short-sighted view, because in the long term not only the paper counts matter, but also one's general reputation within the community.