It's a tad bit unknown, in that article they said one of their primary goals in the first phase of testing is to check for ease of circumvention.
In theory:
It would be possible to make proxies irrelevant. Considering the sacrifices in speed they are willing to put, they could resort to packet sniffing in which case a proxy wouldnt be enough unless you encrypt end to end with the proxy.
But even if normal proxies do work to bypass their content blacklist, they can just blacklist the known proxies. Yes some people would still be able to get out with private proxies or not-as-known proxies, but the vast majority of people would still be locked in. We can assume that if a good proxy becomes known enough to be a problem, then the government will know about it too and be able to block it. They could also in theory blockade means for people to know proxies, Pages with proxy listings, etc.
Like i said, it all comes down (as usual) to how many resources the government wishes to spend in terms of extra equipment to do the sniffing and traffic shaping plus manpower to research appropriate blacklist contents, react as fast as possible to changes, etc. If this just a political move then its doubtful their blockade will do much.
in practice:
Of course, in practice its almost impossible they can do an effective blockade to a motivated person but still, they can probably block average users and thats bad enough. And worst of all, the direction this shows.
When a necessity like this appears, companies like Cisco start investing into R&D to cover it. Just like p2p made traffic shaping popular enough that Cisco eventually made it possible to filter and cap p2p with hardware cheap enough that ISP's in countries like Uruguay can afford it. This is extremely relevant since a hardware solution is what makes it fast enough so its possible to apply in practice.