rt.representation theory - Is Lusztig's conjecture solved?

Geordie Williamson has found counterexamples to Lusztig's conjecture:
http://people.mpim-bonn.mpg.de/geordie/Torsion.pdf

Apparently, any bound on the largest prime that breaks Lusztig's conjecture must be at least proportional to $nlog n$, and this is probably still far too optimistic; I think it's now believed that you'll need an exponential bound.