It's a really interesting paper, but I had two questions about the strong lensing part:
1) Is it remotely plausible without time delays ~100 years?
2) I was confused when you mentioned the lyman alpha forest adding to the signal - does it matter that the multiple images have travelled through slightly different regions? What if the lyman alpha absorbers are rotating?

Thanks Eric. How bright do you think the lensed quasars need to be? How does the trade off between flux and timedelay work out? I looked into redshift drift a few years ago, and based on the assumptions of Liske 2008's paper it seemed that you needed ultra-bright QSOs to make the experiment work.

I guess my long winded question is; do you have an idea of how many suitable lensed targets there are per sky? And if source flux is important, should we have already found all the suitable candidates in the SDSS footprint?

Higuchi et al. want to detect the signal of voids in large weak lensing data. They use N body simulations, and a clever void finder to make a huge stack of voids; this gives a detectable weak lensing signal but only at 5sigma significance for 5000 sqdegree (i.e. the void effect is small compared to other noise in the weak lensing reconstruction [or possibly the void model used isn't very good, given the number of voids stacked]). They find that voids are surrounded by over-dense ridges that have an important effect on the weak lensing signal due to voids. They forecast a 3 sigma detection for an upcoming HSC weak lensing survey.

The void finder looks like it could be a useful tool for people trying to reconstruct the weak lensing convergence/magnification along specific lines of sight. (Although I suspect we still need a better understanding of the outskirts of halos, before we should worry about voids)