I had a dream, quite a few years ago now, that I was standing in an underground parking lot and conversing with an old Astronomy lecturer of mine, Johannes Wolterbeek-Muller.
We were talking about Physics, Astronomy and Computer Programming. While an occasional car passed us, coming in or going out, we considered the uses for programming in the hard sciences. I was convinced – and convincing him, too – that programming was not only extremely useful to Astrophysics, but also that it was somehow of more intrinsic worth than we had thought up to now.
Last year, Patrick Hayden of Stanford started working out how many computational steps would be needed to decode all the information coming out of a black hole, and his results were very interesting.
Leonard Susskind has now taken these thoughts a bit further.
It seems that the computational problem is effectively unsolveable.
There would need to be an exponentially increasing number of steps taken in the program to decode this information.
This result negates, on the face of it, the need for there to be any kind of physical barrier between the outside and the inside of a black hole. The problem is uncalculable – an effective barrier to physical particles both sides of the event horizon being known about and measured.
In addition, the complexity of the information-decoding problem increases exponentially the closer one(the calculating one) gets to the boundary.
Calculation complexity seems to be behaving a bit like gravity. And also like time.
This is a fascinating first pass at an understanding which may – to use a hackneyed phrase – “revolutionise our understanding of the Cosmos”. And also change the ways in which we apply computing to science.