Yes Chaitin is rightly famous for his work - fascinating conclusion about mathematics being an experimental science. Thanks for posting that ..

One of the big challenges facing computational General Rel is how very limited computer programs are in 'real life: at this stage the procedure is to take a 'simple' situation in which an exact mathematical solution to time evolution is already known - such as a Kerr black hole. Then various computational strategies are applied to try and evolve a model of this, over time, and see how these compare to the exact, known solution.

Although they seem to work well with limited parameters, again and again those computer models go haywire over time and end up nowhere near what we know should be the case, from the pen&paper solution. This is the challenge facing numerical relativists and to be frank, thank goodness I had the nous to avoid this valuable but dismal exercise, and choose something more 'artistic'.