jbott

Shouldn't the PCIe bus fix that problem?

I would assume the reason why usb converters do not work would be due to the fact that usb uses a host controlled polling system, resulting in discrete timing that can only be as precise as the bus speed.

One of the authors has additional information posted here: http://pjreddie.com/darknet/yolo/

Yup, this was pretty disappointing to us.

We were keeping it running as long as we felt comfortable to do so, but due to our lack of preparation for an actual attack, we decided to cut it when it was obvious that the majority was voting too perfectly on actions that were turning malicious.

At this point, we are not sure how we are going to be continuing with this project. The time investment required to make this secure is much larger than we initially anticipated and our current setup is not optimal to do so. Along with this, we are both currently students and do not have the time to invest in such an undertaking. However, we are currently talking with a group that is attempting to reboot this idea immediately in a more secure environment. We will be exploring our options on how to best keep this project going.

All of our code is available on github at https://github.com/twitchinstallsarchlinux

No, but we are hoping the chat doesn't die down that much. We considered it, but ultimately didn't consider it necessary.

I have no idea. I don't think the crowd will have enough coordination to do something like that (which is basically what we are banking on currently), but if it does, we will probably have to step in manually.

It works by keypress, which should increase the challenge level (I hope). Combinations like ctrl-c will be one keypress.

Yup! Low delay is awesome. We are getting about a 10 second delay. Save states aren't in quite yet, but that's our next goal.

The problem is that it is possible to put the machine into a state where it does not accept key inputs, so the only way to get it out is to do a "hard reset". We are still not sure how to handle that (Possible a confirmation box where they have to type "yes" to do it?). That whole subsystem is still undergoing revisions, so we are going to do some testing later and see if we can make that less easily exploited.

Hey guys, author here. Feel free to ask any questions, and I'll try my best to answer them!