Shoe Spam

Got this wonderful bit of spam from someone using the name “Red Bottom Shoes”:

a 22-capacity unit out way sailed elation. Rivers complete 20 of 27 for 340 yards, and two contest appearances could formally union Roc. I can sort of the principle can it. It was herded into the gear bracing. I’m not acquiring forwards originally the period of time for Nov, was rated No. Celine Outlet Ray Ban Sunglasses cheap ray ban sunglasses Coach Outlet Online out of his carrying out. It went truly well. He is avid selfish person of or the bodily property into subject of the perfervid tendency to reunify is kind a parent. I’m not into recruiting rankings, O’Brien same.

He battled his give chase for him, same. After nigh two weeks.

The Eagles

Email URL Aliases

Apparently, my Stevens email has 61 different aliases – I only knew about 3 of them. Basically, they are all variations on my email, but with a lot of weird URLs that I didn’t know Stevens owned. They are:

“Sinistral & Dextral” Screenshots




Some screenshots from “Sinistral & Dextral,” the game I built for the Omaha edition of Games++ 2016 under the theme double. In the game, the player navigates a large, procedurally-generated landscape that mirrors itself at various scales (land-forms, rivers, etc mirror across the entire map; trees, rocks, and smaller objects mirror locally). Various “noise patches” populate the map, as do wandering creatures that cause the player to respawn randomly on the map.

The layout of the map, as well as all the colors, are randomized at startup.

(The title of the game refers to the scientific term for “handedness,” a kind of mirrored doubling.)

Torch-rnn: Mac Install


There have been many recent examples of neural networks making interesting content after the algorithm has been fed input data and “learned” about it. Many of these, Google’s Deep Dream being the most well-covered, use and generate images, but what about text? This tutorial will show you how to install Torch-rnn, a set of recurrent neural network tools for character-based (ie: single letter) learning and output – it’s written by Justin Johnson, who deserves a huge “thanks!” for this tool.

The details about how all this works are complex and quite technical, but in short we train our neural network character-by-character, instead of with words like a Markov chain might. It learns what letters are most likely to come after others, and the text is generated the same way. One might think this would output random character soup, but the results are startlingly coherent, even more so than more traditional Markov output.

Torch-rnn is built on Torch, a set of scientific computing tools for the programming language Lua, which lets us take advantage of the GPU, using CUDA or OpenCL to accelerate the training process. Training can take a very long time, especially with large data sets, so the GPU acceleration is a big plus.

Continue reading Torch-rnn: Mac Install

Site Defacement Code

Screen Shot 2016-03-18 at 10.22.32 AM

My server got hacked last night, though my hosting company immediately quarantined the files so no harm was done. But the resulting files are really kind of beautiful from an aesthetic and code-studies level. Below is the “forensic log file” from my host, a little ASCII text file generated by some programmer with a few lines of code listing the suspect file as zebi.php. It actually missed a few other malicious files: cp.php, xp.php, bassimo.php, and bassimo1.php.

The files themselves are obfuscated PHP, but a few hints are left behind. One of the files includes a credit line for FOPO (Free Online PHP Obfuscator) and lists the date/time for obfuscation as Thursday, February 11th, 2016 at 18:47 UTC. It also lists the IP address as A quick whois lookup shows the location of the hackers is (probably/possibly) in Tunisia, and that they use the service provider Agence Tunisienne Internet.

What does it do? I believe it defaces a site, something like this:


Most of the code is PHP, with some HTML, CSS, and a little JavaScript. One notable bit of JavaScript is this one, hosted on, a Turkish hacker site. This script (I think) generates an image of the page it is hacking by passing the URL to a PHP script:

I won’t post the full files here, since they are malware, after all (and readily found online if you want them). But here are a few excerpted screenshots that show a real abstract beauty.

Screen Shot 2016-03-18 at 10.20.53 AM

Screen Shot 2016-03-18 at 10.21.04 AM

Screen Shot 2016-03-18 at 10.23.02 AM

Screen Shot 2016-03-18 at 10.23.21 AM

Screen Shot 2016-03-18 at 10.23.36 AM

Screen Shot 2016-03-18 at 10.23.40 AM

Some of the code above was run through PHP Formatter for cleaning up indentation to more clearly view the code. For a more forensic look at how the code works, try UnPHP for decoding obfuscated code.