More #rC3 talk notes

Andy Müller-Maguhn starts with some disclaimer. This is a talk about how the CIA targeted WikiLeaks collaborators and shows some evidence how Andy himself was surveilled. This talk is necessarily biased and he tries to stay as objective as possible.

First, Andy reviews the CCC’s involvement into WikiLeaks (Wau Holland Foundation collects donations) and the ideology of CCC members shared regarding freedom of information. Andy himself visited Julian Assange in the embassy several times. Then he explains Mike Pompeo’s position towards WikiLeaks that he publicly voiced (Mike Pompeo was a CIA director): “WikiLeaks walks like a hostile intelligence service and talks like a hostile like a hostile intelligence service” and “It is time to call out WikiLeaks for what it really is – a non-state hostile intelligence service”.

When reflecting on these events, he started to attributed a missing sock to Pompeo’s fault. He concluded that he became hypersensitive and the ongoing situational awareness is stressful. It also seems that this situation is infectious as his friends also start to be more sensitive. He wonders what are possible options for the future? He weighs in on a user’s suggestion to become a farmer or maintain a vineyard.

German Heise article on this talk

Pepijn de Vos had the opportunity to work as intern at Symbiotic EDA. First, he explains generically how hardware specification end up on an FPGA. With Verilog, SystemVerilog, or VHDL as frontend, some tool like Yosys, ODIN-II or Quartus performs synthesis. You end up with a JSON, EDIF, or (E)BLIF file. The next step is routing followed by generating the FPGA bitstream. Either you use Vivado, ISE, or Quartus to generate the bitstream or your tooling splits these steps up. First run “Verilog to Routing” routine or “nextpnr” to get fasm or fazm files. Then some open or close FPGA assembler generates the bitstream.

To get started with an FPGA, the following recommendations are given:

The strategy for fuzzing, he developed, is to …

He continues to explain a “Binary trick” and “Balanced constant-weight code”, which I don’t get (the latter only considers bitstrings with constant Hamming weight, the former Idk), but seem to be heuristic approaches to reduce the number of possible bits to flip. He then continues to point out that fuzzing LUT bits is trivial and routing was his major goal, which was a challenge (manual intervention required). IO buffers, on the other hand, are useless (are very complex). He proceeds to explain the clock tree which is also part of each [good] design and then shows the final tile format: 4 LUT rows at bottom and about 80% MUXs above (80%? that is really a lot).

Project apicula

Larissa Willamowski discusses the on-going conflict between Armenia and Azerbaijan. Nagorno-Karabakh is a South Caucasus region that was not known globally as conflict region in previous years, but this changed in 2020-09. According to Larissa, it is a territorial conflict with an ethnic dimension. The region is recognized as part of Aserbaijan under international law.

In conclusion, this talk was a good overview over the conflict. However, the speaker sometimes built complex, long sentences and the loud noise when taking breath made it more difficult to follow.

Anika talks (in German) about drug trading in its third version. She investigated in Berlin (Germany). What does the versioning scheme mean? Anika consider 1.0 as trading on the street. Version 2.0 was trading using the DarkNet and anonymous payments via Bitcoin. Finally, version 3.0 refers to the use of messenger services. She starts to discuss (dis)advantages of a particular distribution form:

Drugs in DarkWeb are actually only a small fraction of the overall DarkWeb, but have the largest trading volume. Anika proceeds to explain cocaine-taxis (cocaine because it had the highest profit margin). Those taxis can be ordered 24/7 and were known by business cards spread around in clubs. The customer enters the taxi and the deal happens. Then the customer leaves again. The organization was similar to actual companies with drivers organized in shifts. Mainly in the context of organized Arabic clans.

Anika illustrates by screenshots how deals actually work on messaging platforms. Here, Telegram is the most important one. The usability is high, phone availability is almost always given, and E2E encryption is theoretically possible. There are two ways of distributions in case of Telegram (i.e. groups and channels). Allows broad overview among many dealers. Delivery happens by car (similar to Cocaine taxis), but they are typically organized by smaller operators. This is usually faster than pizza delivery. Payments are done locally and by cash. On this platform, customers can also ask for offers. Disadvantages include minimal order prices (typically 50€), higher prices (Cannabis costs typically 10€ per gram, but not others), scammers (deliver white powder but not drugs), contact between payer and payee (even enter the car), not completely anonymous, groups and dealers vanish/change quickly, and everyone can “become” dealer.

On instagram, the structure looks a bit different. Dealers contact customers and organize themselves through hashtags. A particular interest is put on younger audience.

What about consequences? Instagram and facebook do not act. Telegram is actively supporting censorship and control (thus a new home of right-wing extremists, drug dealers and conspiracy theorists). Why does the police not just order and catch the drivers of delivery services? Because the police is actively engaged in finding the management people; not the small fish like local distributors. The most popular substances are Cannabis, followed by Cocaine, and party drugs. She concludes: