# Overheard at New Directions in Foundations of Physics ’10

If you missed New Directions in Foundations of Physics conference earlier this month, here are a few memorable one-liners. You’ll notice that many are from Bill Unruh, who seemed to be in rare form that weekend.

In this case it was more important to be right than to be correct. — Bill Unruh

Unruh was discussing Hawking’s original calculation of the thermal radiation emitted by black holes. He ended with a very interesting discussion of “dumb-holes,” along the lines discussed recently on this blog.

Your transparencies should never be more clear than your thoughts. — Bill Unruh

Here, Unruh was struggling to focus one of those new overhead-projectors, the kind that project an actual live video of your transparency. He had just explained his intention to present us with a rough-and-ready example, so the timing was impeccable.

Yes, but… but still, I… I mean, I really shouldn’t tell *you* that you can’t move faster than light. I’m teaching my grandmother to suck eggs, here. — Bill Unruh, to Charlie Misner

As I recall, Unruh had mentioned that you couldn’t escape a black hole unless you traveled faster than light. Misner was gently remindung Unruh that, after all, we have little empirical evidence about the interior of black holes.

Unruh: But those are just words!

Tumulka: Yeah… I don’t know how to convey it better than with words

Here, Roderich Tumulka was explaining some feature of his relativistic GRW theory. I wish I could recall exactly what provoked Unruh to say this, but alas, my memory falters. At any rate, Tumulka admirably deflected his heckling.

I like to say that Feynman won the Nobel prize in 1965 for showing that BQP is contained in PSPACE. — Scott Aaronson

According to a standard picture, BQP is the class of computations quantum computer can do, and PSPACE is just the class of computations solvable in polynomial space. Aaronson was pointing out that, if we use the Feynmann path-integral approach to calculating the probability that a quantum computer “accepts,” then we’re required to sum up only exponentially many terms — and this sum is computable in PSPACE.

If we were physicists, we would have announced decades ago our

discoverythat all these classes [like BQP, PSPACE, NP, etc.] are distinct. — Scott Aaronson

Aaronson had just finished explaining how questions like the distinction between P and NP remain a major open problem in mathematics and computer science. Apparently, he considers the standard of “discovery” in physics to be a notch lower.

Hopefully I’ll see some of you at one of the remaining conferences this summer!

*Soul Physics is authored by Bryan W. Roberts. Thanks for subscribing.*

Want more Soul Physics? Try the Soul Physics Tweet.

Want more Soul Physics? Try the Soul Physics Tweet.

- An argument for hidden variables
- Even more philosophy of physics conferences, Summer 2010

One final note on this extremely intsieetrng topic over on Shtetl Optimized, Scott Aaronson has posted the slides for his QIP talk on CCLs.These notes give an example to show that yes, as discussed in the above posts, the CCL computational output is (in general) a non-smooth function of the computational input.Scott’s notes go on to assert (the mathematical details are not given, but the assertion is highly plausible) that there are some PSPACE-complete unitaries U for which the computational output is a continuous function of the input.My own take on this (from an engineering perspective) is simple:It is commonplace for the state-space trajectories of classical dynamical systems to be non-continuous functions of their initial conditions. When CCLs are added to quantum mechanics, then their state-space trajectories similarly becomes non-continuous functions of the starting conditions; this is (yet another) respect in which CCLs cause quantum systems to resemble classical systems.My own interest in these phenomena is highly practical: these same questions arise naturally, not from CCLs, but from projecting the equations of linear quantum mechanics onto the curved state-spaces of Kc3a4hlerian tensor network manifolds.The point being not that Nature does this projection, but that ordinary human beings do it, within the context of (almost all) known methods for efficient simulation of quantum dynamics. So if there are inherent aspects of the resulting (simulated) quantum physics that are unexpected, we quantum systems engineers definitely want to know about it!This is yet another respect in which even the most arcane aspects of QIT are turning out to have surprising (and important) practical applications.