On Tue, Jun 16, 2015 at 5:26 PM, Sean Lynch <[1]seanl@literati.org>
   wrote:

     Lots of words, very few details. Fonts getting a "bit pixellated"?
     Are you kidding me?

   [2]http://www.pcworld.com/article/2921092/gpu-malware-can-also-affect-w
   indows-pcs-possibly-macs.html
   There's an endless number of ways that malware that don't follow any
   neat process isolation model with clean usage of API:s can cause what
   would be experienced as glitches. Hiding executables in GPU memory
   assigned to fonts can do that. And yes, that's 100% possible.

     Packages "piggybacking on other packages"? This is all very
     imprecise language for someone who is attempting to convince us that
     something very grave is going on. And as usual, not a single hex
     dump of a single packet. Not of any of the packets supposedly
     spewing out of their supposedly disabled Ethernet port, not out of
     their supposedly disabled wifi card, not of one of these supposedly
     piggybacked packages.

   They might not want to show examples of the injection attacks in order
   to not reveal how they're detecting the traffic. Look up NSA's Turmoil
   and Quantum Insert.

     I'm not saying these capabilities don't exist; I'm sure they do. I'm
     not even saying the author is lying or stupid.

   First of all, it is written mostly for a non-technical audience.
   Second, you're a bit stuck on the high-level models of computers here,
   you're not considering how the effects of binary level tampering and
   code exploits and altering RAM and even firmware for persistence
   attacks ([3]http://www.wired.com/2015/02/nsa-firmware-hacking/) might
   manifest themselves. To somebody who thought he really did secure his
   systems well, the signs that well obscured malware will show will make
   it look like your computer has ghosts.

     Even assuming some of these claims are true, not asking for more
     evidence robs us of the ability to defend ourselves. Running off to
     build f2f networks is fun and all, but it's not going to do a lick
     of good if we have no idea what we're up against beyond some vague
     descriptions, especially when you consider that the capabilities of
     our adversaries go well beyond the technological. There is such a
     thing as technological security that's "too good", when you've spent
     all your time defending against technological attacks only to
     succomb to, as others on this thread have pointed out, a rubber
     hose.

   One problem is that the attacks change too fast. Holding off until they
   change it themselves can allow you to detect even more of their
   activity than anybody would if you told the world right away. Otherwise
   they'll instantly stop the particular attacks you detected and switch
   to something else. At best one could release details of how to analyze
   your old offline backups for signs of infections.

References

   1. mailto:seanl@literati.org
   2. http://www.pcworld.com/article/2921092/gpu-malware-can-also-affect-windows-pcs-possibly-macs.html
   3. http://www.wired.com/2015/02/nsa-firmware-hacking/