[ExI] suicide pilots

[Anders Sandberg]

William Flynn Wallace <foozler83 at gmail.com> , 26/4/2015 6:13 PM:



Isn't everyone trying to hack into everything, military or not?  Satellites, missiles, drones and more I assume.  I hope we can hack into drones when they are used against us, as they inevitably will be.

If the military can have a secure system for , say, satellites, why can't the airlines?  Will we never have a secure system of anything?  Is that our future?​ 

That is actually a very good question. 


Teleoperated surgery looks quite hackable: http://www.technologyreview.com/view/537001/security-experts-hack-teleoperated-surgical-robot/ but one could argue that what has happened is that the medical tech world are just late in realizing the need of software security. Just like the lab automation world. And credit card terminals: http://boingboing.net/2015/04/24/credit-card-swipe-machines-hav.html ...and so on.


I think the deep question is whether the attacker will have the advantage, or the defender. A common theory among computer security experts is that it is attacker advantage, in which case we will rarely have secure systems. Sure, it is possible to design provably secure systems, but they are limited, expensive or not as enticing as the open and insecure systems, so the latter will dominate. But the theory is by no means proven (or even provable?)


My own answer is to look at analogies. From one of my papers: "Self-propagating parasitical replicators are common in open-ended information processing systems and perhaps extra common when copying is easy:

• Transposable elements, sequences of DNA that can move or copy themselves, make up a large fraction of eukaryote genomes (55% in humans, 60% in maize (Kidwell 2002)). In addition, there are many other forms of intragenomic conflict where genes have evolved to propagate themselves to the detriment of other genes or their hosts.
• Viruses appear to be the most numerically abundant replicators on Earth (Suttle 2005, Steward et al. 2013). Given that they seem to be about an order of magnitude more numerous than prokaryotes, this means that a randomly encountered genome is ≈90% likely to be viral. 
• Parasites can make up a substantial fraction of the biomass in ecosystems (Preston et al. 2013), sometimes even outweighing their hosts (Kuris et al. 2008). 
• More than 75% of email messages are spam, with about 7% containing malware (Microsoft 2013). It may be that the release rate of malicious software exceeds legitimate software (Symantec 2008) (although biased reporting and the difficulty of determining what strains to count separately should make one take the claim with a bit of salt). "


One can also argue that defence budgets has a median of about 2.3% of GDP. Policing budgets are typically about 1% of GDP. We spend 10% of our metabolism on the immune system (and our resting metabolic rate goes up by more several tens of percent when we have an infection - in some case *way* more, like 300% for severe cases). 


Taken together, this suggests that we should expect that outside systems made so brittle or solid that hacking is not possible, there will be enough of it that it warrants using a fairly noticeable fraction of the total utility for defending them. 




Anders Sandberg, Future of Humanity Institute Philosophy Faculty of Oxford University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20150427/f5007a69/attachment.html>