I //have// read it. If Friendly AI comes along, we have nothing to talk about now. I have plenty to say on that topic (and I need to re-read the material one more time), but it's off topic here.
Moving on...
I should make clear that I'm conflicted on the centralized factories issue. In thinking about this, I think it's important not just to visualize what we can or should do there, but to consider the fact that we have to get there //from here//. Here, in this case, is an American political system which largely overrun by industrial interests and irrational fear of terrorism.
The security issue is, I think, more than enough to justify public, central factories. That's a normative policy recommendation on my part (as much as I hate to espouse it). But there is also the issue of protecting the overbearing intellectual property rights of private industrial interests. If Hollywood has enough political influence to bribe politicians and essentially buy legislation that favors their market dominance (Sonny Bono Act, Digital Millenium Copyright Act, etc), then the Manufacturing sector will be orders of magnitude more powerful, because their revenues and employment numbers frankly DWARF the copyright cartel. Whenever nanofactories (or even macroscale recursive fabrication) begins to mount a significant threat to their marketshare, they will have the political influence they need to retain a lock on them. In addition, the image of the nano-terrorist boogey-man, amplified by our present irrational post-9/11 fearmongering national culture, will allow them to play the system RIGHT into their hands. So this isn't just a (reluctant) policy recommendation. It's also a prediction.
Aside from all this, the principle of accountability is what leads to the idea of production being a public act. After some thought, it is possible to imagine how one could apply that same model to desktop nanofactories. You begin to suggest this yourself, where the nanofactory keeps production records and energy budget accounts, etc. I compare that with the publically-surveilled centralized facilities, which are under surveillance 24/7, //not just when the nanofacs are being used//. So why not throw video surveillance on the desktop units, which would then keep track of what their users are up to around it? Of course, this would mean that the factories themselves are not suitable for use in the average private home - because, well, your privacy is gone. And that's the point. They may be portable and avaialable enough to cart into private spaces, but the fact is they're to dangerous to allow people //to use them privately//.
Another reason to do this is that, if you're insisting that people take home a black box that they are prohibited from knowing much about, you introduce real dangers of the creators hiding other kinds of nasty things in it in secret. Like, oh .... surveillance cameras! Meaning, cameras that only certain people have access to. By intentioanlly putting cameras into them //and announcing// their presence, people won't be lulled into a false sense of privacy. This problem is similar to that of
voting machines.
"I'm not assuming that in this paper because you can only take on so many unpopular positions before people start labeling you a cultist."
You'd be surprised what people
are getting away with these days.
You haven't addressed my objection to the computational speed limit issue. How do you propose to stop people from clustering slower, unrestricted computers to acheive the same capabilities?
Since you bring up quantum computers, I should point out that quantum computation appears to
nullify all public key encryption applications, which has disastrous consequences not only for Trusted Manufacturing DRM, but also for Trusted Computing DRM, privacy, ecommerce... the works. My reasearch has turned up theoretical work that shows how QC can be applied to break every know asymmetric algorithm (ie, anything that doesn't require a single key), and none to the contrary.
As far as "
quantum encryption" goes, it really isn't encryption as such. Quantum //Key Distribution//, aka Quantum key exchange, can be used to implement a subset of a cryptosystem involving the transmission of a symmetric key. Today, this is normally done by asymmetric encryption algorithms. The Achilles' heel of QKD is that it can't be done over the Internet; you have to have a complete uninterrupted optical circuit between sender and receiver, implying that a circuit switched fiberoptic network or a direct laser or microwave wireless link would be necessary. "quantum encryption" will do nothing to protect data already stored within the nanofactory - it protects communications.
It's interesting to talk about the intended effects of Intellectual Property. The US
constitutional mandate for IP is intended to encourage innovation. Recently, there's been
some extensive research published showing how the opposite is nearly always the case. Ironically, however, not only do we have to question whether or not IP licensing regimes really encourage or retard innovation - but we have to ask ourselves //which of those two outcomes we should favor//. The Constitution says we should speed it up. In that sense, there's a mandate to support the (Kurzweilian) singularity right there in the Constitution! Aside from that, though, we have to consider that //slowing it down// might be a prudent idea, and, perversely, the possibility that an Intellectual Property regime might just be the thing to do it! This is totally bizaare, but (or so) I thought I'd mention it.
""Monopoly" is this scary big word that people use, but does it bother us that Google has an almost-monopoly on search?"
YES, it does. Voicing privacy concerns is not just an obscure hobby practiced by marginalized loons, as your response suggests. As I said, if you're going to permit a monopoly anyway, in order to leverage the numerous advantages of centralization, strict government regulations, and technical standards, you might as well subject it more directly to democratic control (in theory, anyway
sigh) by socializing the IP, design, R&D, etc., and making it a civic, rather than a private institution.
Have you any references to any serious engineering work done on tamper-proofing nanofactories, or is this just what you've imagined so far?
From here on out, I find us in agreement, except for the last paragraph. I'll leave that for some other time, though.
Thanks for your work, on this, Michael. It's fantastic to get people discussing these very important issues.
--Nato
