i would like to know more about the bolded
General idea, some of the things that make it so dangerous are also what make it fragile.
One of the things about HIV is its initial entry into a host. In most cases, it seems to be dependent on interactions with dendritic cells or DCs (Technically iDCs. IV drug abusers probably allow the virus to skip this step). There's a specific receptor on DCs that mediates this action. Change the outer surface of HIV significantly, and you probably eliminate this interaction, and thus maybe a lot of infections.
Furthermore, the main target on HIV for the immune system is a molecule called GP120 (a simplification but true at least for antibodies). It's a very special glycoprotein in many ways. There's a lot to it, but two things in particular. First, it's why CD4 T cells are the main target of the virus. Second, it seems to have a very unusual property in that it can be 'attacked' (I hate that word, but I'll use it here) by antibodies and still continue to function more or less normally. That's
extremely unusual, maybe unique.
If GP120 and the general envelope structure change too much (i.e. towards a more stable phenotype) you're much more likely than not to change these properties. It wouldn't target T cells (or any cells?) so efficiently and it would be much more susceptible to an immune response.
bad analogy: if we removed the wheels from a car, it would no longer get flat tires. But it's not really much use as a car anymore.