Yes, yes, there are problems with the initiative
because many researchers currently depend on
non-open-source software; there are other issues
On the other hand, there are very serious ethical
issues with the status quo: professors paying grad
students minimum wage to do the work, then using
it to start a company. (Many universities claim
that their research assistants work 20 hours a
week when they live in the lab, putting in 60
hours/week+, especially on cutting-edge, potentially lucrative research, hence their stipends are usually around minimum wage).
At minimum, universities receiving taxpayer money
should not be permitted to prevent researchers
who wish to release their research software as
open source from doing so in cases where this is
legal (that is, the software in question is not
a derivative work of some non-open-source software).
A university might still be able to recover some
revenue by charging proprietary software developers
for the right to incorporate the code with fewer
restrictions, e.g. GPL for the general public,
more permissive licensing for those willing to pay.