Adobe Neural Net Detects Photoshop Shenanigans

June 18, 2019 by Al Williams 13 Comments

Photoshop can take a bad picture and make it look better. But it can also take a picture of you smiling and make it into a picture of your frowning. Altering images and video can of course be benign, but it can also have nefarious purposes. Adobe teamed up with researchers at Berkeley to see if they could teach a computer to detect a very specific type of photo manipulation. As it turns out, they could.

Using a Photoshop feature called face-aware liquify, slightly more than half of the people tested could tell which picture was the original and which was retouched to alter the facial expression. However, after sufficient training, Adobe’s neural network could solve the puzzle correctly 99% of the time.

Continue reading “Adobe Neural Net Detects Photoshop Shenanigans” →

Exploring The Dell N1108T-ON Ethernet Switch

June 17, 2019 by Tom Nardi 24 Comments

In an era where everything seems to be getting “smarter” every year, it will probably come as no surprise to find that even relatively middling networking hardware is now packing advanced features and considerable computational power. A case in point is the Dell N1108T-ON Ethernet switch. Despite only costing around $100 USD on the second hand market, [Ben Cox] discovered this particular switch was capable of a lot more than what was advertised by poking around its onboard operating system.

It all started by plugging into the serial port on the front of the switch, which [Ben] happily notes is an integrated FTDI USB serial adapter to make life easy. Booting into recovery mode gave him local shell access, and some poking around determines it’s the sort of BusyBox-powered Linux system that you’d expect on an embedded device. The biggest discoveries were that it was running a relatively recent kernel (3.8.1), and that it apparently had Python installed.

From there, [Ben] found out that these switches have a feature where the administrator can install and run Python “applications” by packaging them up as tarballs and copying them from a USB flash drive. So he wrote up a simple Python program that used the socket library to open up a reverse shell to his desktop computer, and to his surprise, it worked perfectly on the first try. Now with root access, the fun really started.

The next step was getting an SSH installed and running on the switch, so that he didn’t have to do the reverse shell trick every time. He then started installing the packages necessary to turn the switch into a secure VPN tunnel with Wireguard. This took a little fiddling as [Ben] didn’t have the option of installing the normal Wireguard kernel module, but he eventually got the necessary tools modified and cross-compiled to ARM. He believes this is just the start of what’s capable on devices like this, and we’re interested in seeing where the community goes from here.

We’ve seen hackers add management capability to a “dumb” unmanaged switch in the past, but software modifications like this promise to make the creation of custom, secure, networks far easier even on a hacker’s budget. A lot has certainly changed since the last time we saw somebody really dive into a professional Ethernet switch.

a 3d mesh of a rabbit, and a knit version of the same

Knitting Software Automatically Converts 3D Models Into Machine-knit Stuffies

June 14, 2019 by Brian Boucheron 20 Comments

We’ve seen our fair share of interesting knitting hacks here at Hackaday. There has been a lot of creative space explored while mashing computers into knitting machines and vice versa, but for the most part the resulting knit goods all tend to be a bit… two-dimensional. The mechanical reality of knitting and hobbyist-level knitting machines just tends to lend itself to working with a simple grid of pixels in a flat plane.

However, a team at the [Carnegie Mellon Textiles Lab] have been taking the world of computer-controlled knitting from two dimensions to three, with software that can create knitting patterns for most any 3D model you feed it. Think of it like your standard 3D printing slicer software, except instead of simple layers of thermoplastics it generates complex multi-dimensional chains of knits and purls with yarn and 100% stuffing infill.

The details are discussed and very well illustrated in their paper entitled Automatic Machine Knitting of 3D Meshes and a video (unfortunately not embeddable) shows the software interface in action, along with some of the stuffing process and the final adorable (ok they’re a little creepy too) stuffed shapes.

Since the publication of their paper, [the Textiles Lab] has also released an open-source version of their autoknit software on GitHub. Although the compilation and installation steps look non-trivial, the actual interface seems approachable by a dedicated hobbyist. Anyone comfortable with 3D slicer software should be able to load a model, define the two seams necessary to close the shape, which will need to be manually sewn after stuffing, and output the knitting machine code.

Previous knits: the Knit Universe, Bike-driven Scarf Knitter, Knitted Circuit Board.

Hyperlinking Comes To GitHub Via Extension

June 14, 2019 by Al Williams 8 Comments

If you are browsing GitHub it is very tempting to open up the source code to some project and peek at how it works. The code view is easy to read, but the viewer lacks one important feature: the ability to click on an included file and find it. The Octolinker extension fixes that oversight.

If you want to try it without installing the extension, there is a mock-up demo available. Even though the demo wants you to click on specific things, if you don’t play by the rules it will still do the right thing and take you to either the code on GitHub or an appropriate page. You can even substitute the demo URL for github.com and try it out on any GitHub page without the extension.

Continue reading “Hyperlinking Comes To GitHub Via Extension” →

Windows 10 Goes To Shell

June 10, 2019 by Al Williams 51 Comments

Windows 10 — the operating system people love to hate or hate to love. Even if you’re a Linux die-hard, it is a fair bet that your workplace uses it and that you have friends and family members that need help forcing you to use Windows at least some times. If you prefer a command line — or even just find a place where you have to use the command line, you might find the classic Windows shell a bit anemic. Some of that’s the shell’s fault, but some of it is the Windows console which is — sort of — the terminal program that runs various Windows text-based programs. If you have the creator update channel on Windows 10, though, there have been some recent improvements to the console and the Linux system that will eventually trickle down to the mainstream users.

What’s New?

So what’s new? According to Microsoft, they’ve improved the call interface to make the following things work correctly (along with “many others”):

Core tools: apt, sed, grep, awk, top, tmux, ssh, scp, etc.
Shells: Bash, zsh, fish, etc.
Dev tools: vim, emacs, nano, git, gdb, etc.
Languages & platforms: Node.js & npm, Ruby & Gems, Java & Maven, Python & Pip, C/C++, C# &
.NET Core & Nuget, Go, Rust, Haskell, Elixir/Erlang, etc.
Systems & Services: sshd, Apache, lighttpd, nginx, MySQL, PostgreSQL

The changes to the console are mostly surrounding escape sequences, colors, and mouse support. The API changes included things like allowing certain non-administrative users to create symlinks. We’ve made X Windows work with Windows (using a third-party X server) and Microsoft acknowledges that it has been done. However, they still don’t support it officially.

Continue reading “Windows 10 Goes To Shell” →

3D Printering: The Past And Future Of Prusa’s Slicer

May 24, 2019 by Tom Nardi 41 Comments

If you own a desktop 3D printer, you’re almost certainly familiar with Slic3r. Even if the name doesn’t ring a bell, there’s an excellent chance that a program you’ve used to convert STLs into the G-code your printer can understand was using Slic3r behind the scenes in some capacity. While there have been the occasional challengers, Slic3r has remained one of the most widely used open source slicers for the better part of a decade. While some might argue that proprietary slicers have pulled ahead in some respects, it’s hard to beat free.

So when Josef Prusa announced his team’s fork of Slic3r back in 2016, it wasn’t exactly a shock. The company wanted to offer a slicer optimized for their line of 3D printers, and being big proponents of open source, it made sense they would lean heavily on what was already available in the community. The result was the aptly named “Slic3r Prusa Edition”, or as it came to be known, Slic3r PE.

Ostensibly the fork enabled Prusa to fine tune print parameters for their particular machines and implement support for products such as their Multi-Material Upgrade, but it didn’t take long for Prusa’s developers to start fixing and improving core Slic3r functionality. As both projects were released under the GNU Affero General Public License v3.0, any and all of these improvements could be backported to the original Slic3r; but doing so would take considerable time and effort, something that’s always in short supply with community developed projects.

Since Slic3r PE still produced standard G-code that any 3D printer could use, soon people started using it with their non-Prusa printers simply because it had more features. But this served only to further blur the line between the two projects, especially for new users. When issues arose, it could be hard to determine who should take responsibility for it. All the while, the gap between the two projects continued to widen.

With a new release on the horizon that promised to bring massive changes to Slic3r PE, Josef Prusa decided things had reached a tipping point. In a recent blog post, he announced that as of version 2.0, their slicer would henceforth be known as PrusaSlicer. Let’s take a look at this new slicer, and find out what it took to finally separate these two projects.

Continue reading “3D Printering: The Past And Future Of Prusa’s Slicer” →

Reverse Engineering Dropbox

May 17, 2019 by Al Williams 7 Comments

These days everyone talks about data “in the cloud.” However, before that phrase was fashionable, there were a few pioneers and one of the most famous of these is Dropbox — a service that let you store files on a remote server that dates back to 2007. [Vincent Berg] first noticed some odd network traffic on a hotel network, and figured out that it was a feature of Dropbox that allows computers on the same network to update each other. This led him to start investigating the undocumented Dropbox protocol and reverse engineering the Linux client.

We won’t ask why [Vincent] was poking around the hotel network to start with. However, a cursory glance at the Dropbox client gave away that it was using Python. The byte-compiled classes were — at the time — in a ZIP file added to the executable (which was nothing but a modified copy of Python). The files were encrypted, but [Vincent] used a clever technique. He built a shared object using normal Python and put a backdoor in it that gave him access to the Dropbox Python interpreter.

Continue reading “Reverse Engineering Dropbox” →