You could want a falsifiable prediction about an actual computer. DZ is quite explicit about this: infinite tapes don’t exist. It’s one thing to respond “They do too make sense,” but to respond “They do too exist” is not only wrong, but seems to me to be an unfair response.

That is not to say I can extract a consistent position from DZ. He seems to be willing to quantify over all sizes of machine, which does not seem to me very different from allowing infinite sizes. (At least, it is easy to translate Turing’s results into this language. It is probably possible to formally translate all of TCS, including oracles, into this language, which ought to adequate to define the meaning to DZ’s satisfaction. But some uses of infinity by mathematicians probably cannot be translated. Actually, the random oracle hypothesis probably cannot be so translated.)

For example, $100 would be encoded as a bit string of 1 terabyte. This would be analogous to gold based currency – it’s currency backed by server storage.

Can you come up with a way to encode currency such that it’s backed by the average market cost of CPU time?

• noise is equivalent to measurement,
• measurement suppresses decay,
• the interior of a star is noisy,
• so nuclear lifetimes are longer in stars.

But this is not observed. Why?

And yes, our recent unpublished work, on quantum obfuscation and quantum copy-protection, should have applications to secure cloud computing. However, I should mention that secure cloud computing is in some sense an “easier” problem than obfuscation and copy-protection—indeed, the recent breakthroughs on fully homomorphic encryption let you do secure cloud computing under a plausible classical cryptographic assumption, with no quantum needed. And conversely, the recent breakthroughs on blind quantum computing let you do secure cloud computing with quantum and no cryptographic assumptions. By contrast, to get obfuscation of arbitrary programs, or copy-protection of any programs, you seem to need both quantum and cryptographic assumptions.

I had not heard of this until recently, but there is this idea of a topological quantum computer that is quite different from the “decorherence prone” “ordinary” quantum computer. I’ve never seen you blog about it. And the braid-based logic circuits appear to be quite cool—probably involves some nicer math there.

I would love to learn more about TQC, if you will explain it to me.

Thank you.
SB

At the end of the talk you mentionned your scheme could help protecting softeware copyright. Could it also provide a way to securize cloud computing?

Is it possible that verifying money state is also by itself a procedure to correct for potential noise?