Who said anything about being stuck? And you seem as if you have some kind of resentment for Gl1, not to say that I don't agree with you, it's not particularly the most effective, or technologically advanced system, but it's the framework for all the other resources openGl has to offer. Besides, a good developer could use something like Gl1 and still make a great game, because it's not about making sure it meets the current standards. it's about creating a compelling experience that demands interactivity.
a drawback here is that OpenGL 1.x and newer versions of OpenGL diverge sufficiently WRT how they behave (and what sets of API calls are supported, ...), that one is left essentially needing multiple versions of their renderer:
one to work well on older hardware (and some vaguely new but low-stat HW);
one to use newer features and work on newer hardware;
possibly one to deal with OpenGL ES.
while a person could just do a sole GL 1.x renderer, they may find:
it performs poorly on newer hardware (vs what is possible with a different rendering strategy);
it risks not actually working in some cases (such as those where many of the older API calls were dropped);
this results in portability either requiring an abstraction layer in the renderer (to support alternate targets), or creating a wrapper layer and partially emulating GL 1.x on top of newer or alternate GL versions (crufty and not terribly efficient).