(no subject)

[foxgrrl]

This is what I actually do for a living, seriously:

http://xkcd.com/350/

Speaking of which, I'm (or should be) working on an abstract for a presentation for Security Opus and Black-Hat Europe, next spring. I'm not 100% certain what I'm going to talk about, since I'm still researching it, so it will be sufficiently vague that I can fill in the details… once I have them. Considering how many talks never match up the the vague blurbs in the schedules, I think most presenters do this.

Anyway, what would you, the LJ-reading audience, like to hear me give a presentation on?

[Poll #1096517]

There are some other projects I have too, that I could probably talk about, but I don't think I'll have them done in time.

Comments

[ephermata.livejournal.com]

I voted for a few things, but the QEMU project would kick ass. Bonus points if you can integrate functionality like Valgrind's memcheck to give the world a real kernel memory checking debugger. :)

[ephermata.livejournal.com]

Funny how people work out, isn't it? :) Hope things are going well for you.

[harlkyn.livejournal.com]

omg, you are way way smarter than me :D

[ilcylic.livejournal.com]

*blink blink*

Damn. Y'know, I generally think of myself as fairly well educated, but the hell if any of that made any sense to me at all. Well, except the in-joke about Cmdr. Taco, of course.

[anaisdjuna.livejournal.com]

I voted for the first one. Solving the horizontal gene transfer and the model of the two parallel worlds sounds intriguing.

They all sound hawt tho ;-P

[anaisdjuna.livejournal.com]

oh and i totally thought of you when i saw that xkcd.

wildly cute.

[dragonzuela.livejournal.com]

Whoa, I'm taking a course on phylogenetics, but like, using DNA and protein sequence data. I didn't know people were using it for analyzing malware! Nifty.

[tobyhush.livejournal.com]

If you are looking at using biological methods to look at malware you need to talk to Gary Golumb about what he's been up to. And I suspect that Offensive Computing has done some work in this space also.
The QEMU would be really cool, I seem to recall some similar work/discussion at this year's defcon.
kernel methods for vector machines sounds damned cool.
The shellcode polymorphism is... kinda done to death right now I think? Seems like attackers have defenders solidly beat in terms of detecting shellcode on the wire at least for a while, just with the functions available in metasploit and company.
Wireshark mods would be handy but I'm not sure they are worth a talk.

[dv-girl.livejournal.com]

Sorry I missed you Sunday. The Kaiju battle got shifted to Sunday due to cast issues. I'm free this entire weekend though if you'd like to get together and do something.

[aaron-raccoon.livejournal.com]

I've always thought something like QEMU would make an awesome kernel debugger.

[darkmane.livejournal.com]

Hi. wingedelf mentioned that you did what that xkcd comic was talking about, so I thought I would come over and peer into your life if you don't mind.

I voted for the ideas I understood. I don't know why the world needs another kernel debugger, probably because of my lack of knowledge of strengths and weakness' of existing kernel debuggers.

[saizai.livejournal.com]

Any of the above would be hot.

But make sure you post a video of it so that we can watch. :)