Now these natural reactors were discovered in 1972, but they might have been found decades earlier providing a remarkable proof of concept for a man-made reactor.
Historically, there have been other examples of natural nuclear reactors, e.g. the nearest star, our sun. In the late 19th century, it was realized that the sun and the earth were far older than had been realized, which raised all manner of difficult questions. The known energy reactions at the time would have left the sun a dead cinder after a few thousand years. Since the sun was a good million times older than that, obviously something radically different in the way of energy production/processes was going on, long before fusion reactions in the stars were understood. Once they were understood, of course, even more spectacular sources of stellar energy (e.g. supernovas) began to be understood as well.
The point of all this is that if the conversion of angular momentum <--> linear momentum does take place, it has to do so at the quantum level. Then, if this hypothetical conversion is to have any observable consequences, the stars are the place to look for it, the stars undergoing extreme changes of state shall we say. So far there are no observed exceptions to the conservation of energy with any of the extreme stellar events that have been observed. Despite the fact that short term, out of sight, violations of conservation of energy are permitted and do happen at the quantum level, they are not carried forward into the universe we are familiar with. Which is unfortunate, because as Anastopoulus [PoW, pg. 145] writes: "General Relativity is not characterized by the symmetry of time translation because the notion of time depends on the geometry, and the geometry changes dynamically. For this reason, the concept of energy is not defined in General Relativity." But for all known observations it might as well be.
The same follows for the concept of momentum (angular and linear).
So time translation symmetry holds as far as we can see even given the General Theory of Relativity, which as noted is not a friendly domain for the Poincare space-time symmetries we are familiar with. The geometry of space-time is "not predetermined and absolute like in Special Relativity."
But, and this cannot be emphasized enough, so far no observed deviations from the fundamental conservation laws have been found.
So what is a physicist to do? One approach to seeing if this conversion of angular to linear momentum takes place is to maintain and expand the neutron star research program to get to the bottom of the anomalous neutron star velocities. This is priority number one. A real physicist would say that because a supernova explosion in its early stages results in a number of asymmetrical jets, known physics can count for all unusual observations to date. Maybe so, but the evidence is insufficiently conclusive to me and on that basis I would like the observations and modeling to continue as far as possible for as long as possible. Theory both at the quantum and GR level says the conversion of angular <--> linear momentum, the breaking of the Poincare symmetries, cannot be ruled out. But if the symmetry breaking isn't observed at the most extreme conditions taking place in the universe, i.e. the core of a star going supernova, that would put some serious constraints on the theory of Quantum Gravity. The implications to science are too crucial to do anything less than carry forward on this program.