Saturday, September 10, 2016

Randomness does not need new technology

SciAm reports:
Photonic Chip Could Strengthen Smartphone Encryption

The chip uses pulses of laser light to generate truly random numbers, the basis of encryption. Christopher Intagliata reports.

Random numbers are hugely important for modern computing. They're used to encrypt credit card numbers and emails. To inject randomness into online gaming. And to simulate super complex phenomena, like protein folding or nuclear fission.

But here's the dirty secret: a lot of these so-called random numbers are not truly random. They're actually what’s known as "pseudo random numbers," generated by algorithms. Think of generating random numbers by rolling dice. If you know the number of dice, it’s simple to figure out something about the realm of possible random numbers—thus putting probabilistic limits on the randomness.

But truly random numbers can be generated through quantum mechanical processes. So researchers built a photonic chip — a computer chip that uses photons instead of electrons. The chip has two lasers: one shoots continuously; the other pulses at regular intervals. Each time the two lasers meet, the interference between the light beams is random, thanks to the rules of quantum mechanics. The chip then digitizes that random signal, and voila: a quantum random number generator.
This distinction between truly random and pseudo random numbers is entirely fallacious.

They act as if quantum hanky panky magically makes something more random than other random things.

See the Wikipedia article on Hardware random number generators. There are lots of them on the market already. They are already cheap enuf for smart phones, if there were any commercial need.

I actually think that it would be nice to have such a random number function in PCs and phones. But the fact is that there are suitable workarounds already. There is no need for a fancy 2-laser device like in this research.

Massimo Pigliucci is a philosophy professor who specializes in calling things pseudoscience, and argues strenuously that it is a useful term:
Pseudoscience refers to “any body of knowledge that purports to be scientific or to be supported by science but which fails to comply with the scientific method,” though since there is no such thing as the scientific method, I would rather modify the above to read “with currently accepted scientific standards.” ...

Burke then goes on to cite a 2011 dissertation by Paul Lawrie that argues that “dismissing… scientific racism as ‘pseudo-science,’ or a perversion of the scientific method, blurs our understanding of its role as a tool of racial labor control in modern America.”

Maybe it does, maybe it doesn’t. But scientific racism is a pseudoscience, and it is widely recognized — again by the relevant epistemic community — as such.
It seems to me that the term pseudoscience is mainly used for political purposes and unscientific name-calling, such as for denouncing racism.

But if anything is pseudoscience, then why not the creation of "truly random numbers"? Why is it acceptable to distinguish truly random numbers from not-truly random numbers, but only pseudoscience to distinguish Caucasians from Orientals?

Current philosophers like Pigliucci should be called pseudo-philosophers. They sound as if they are doing philosophy, but they fail to follow minimal standards.

1 comment:

  1. All of this stuff is because academia is a bubble. Security problems have almost nothing to do with encryption and software writers have no skills to avoid common errors such as timing attacks and buffer overruns.

    ReplyDelete