33 karma · joined May 31, 2020
Say you have some computable number p, that means you can compute a (rational) approximation p' to p within any given tolerance eps > 0 (i.e. you know |p - p'| < eps). To determine whether p > 0, p = 0, or p < 0, you compute an approximation p' to a certain tolerance eps. If p' > eps then you know p > 0, if p < -eps then you know p < 0, otherwise you need a better approximation... Without further knowledge about p, there is no point where you can assert p = 0.
I wonder about the use of H1B vs O1 in academia. When I was doing postdocs in the US, I was often told that O1 was not the best route and H1B was easier (edit: for permanent positions). I learned from older colleagues that O1 was actually more frequent in their time (probably in the 90s). This is hearsay, of course, but I heard it often enough that seems somewhat accurate. What changed and could the H1B vs O1 balance change in the future?