biosboiii

7 ms·

31.
by biosboiii 2y ago
Yeah, you can basically just unzip IPA files. Gaining them is hard though, I have a pathway if you are interested. But the Objective C code is actually compiled, and decompilation is a lot harder than with the JVM languages on Android. My n
32.
by biosboiii 2y ago
You're right about the TPM, I won't get the key out of it. It's a special ASIC which doesn't even have the silicon gates to give me the key. But is the TPM doing matrix-mulitiplication at 1.3 Petaflops? Or are you just s
33.
by biosboiii 2y ago
I think you are starting off from the perfect direction, being a forward-engineer first, and then a reverse-engineer. The community around Frida is a a) a bit small and b) a bit unorganized/shadowy. You cannot find that many resources,
34.
by biosboiii 2y ago
Your second note is very interesting, having looked at the model myself this is very plausible. For models which use a lot of input nodes, a lot of "hidden layers" and in the end just perform a softmax this may get infeasible beca
35.
by biosboiii 2y ago
But then you could compromise the GPU, probably :) Look at the bootloader, can you open a console? If not, can you desolder the flash and read the key? If not, can you access the bootloader when the flash is not detected anymore? ... Can yo
36.
by biosboiii 2y ago
Thanks a lot :) My general writing style is directed mainly towards my non-technical colleagues, which I wish to inspire to learn about computers. This is no novelty, by far, it is a pretty standard use-case of Frida. But I think many peopl
37.
by biosboiii 2y ago
Author here, no clue about homeomorphic (or whatever) encryption, what could certainly be done is some sort of encryption of the model into the inference engine. So e.g.: Apple CoreML issues a Public Key, the model is encrypted with that Pu
38.
by biosboiii 2y ago
Author here, it would be nice to claim that I did this on purpose but I really did not know it was open source. I was rather interested in the process of instrumenting of TF to make this "attack" scalable to other apps.
39.
by biosboiii 2y ago
I think the comment author means offering inference via Firebase, with the model never leaving the backend. This works, just like ChatGPT works, but has the downside of 1. You have to pay the computing for every inference 2. Your users can&
40.
by biosboiii 2y ago
Hey, author here. I understand that its not very clear if the neural net and its weights & biases are considered as IP, I personally think that if some OpenAI employee just leaks GPT-4o it isn't magically public domain and everyone
41.
by biosboiii 2y ago
Thanx! Yeah, they should :) Would love to do this with CoreML on Apple devices, but my newest iPhone is a 7. But if you subscribe, you may see me doing the same with a surveillance camera soon(isch) :)
42.
by biosboiii 2y ago
Check out my reverse-engineering/cracking of Microsoft's App just doing that, SeeingAI. https://altayakkus.substack.com/p/you-wouldnt-download-an-ai
43.
You wouldn't download an AI – Stealing mobile AI models with Frida(altayakkus.substack.com)
5 points by biosboiii 2y ago|0 comments
44.
by biosboiii 2y ago
Qualcomm being Qualcomm, again and again and again.
45.
by biosboiii 2y ago
Since MISRA is targetted at Automotive, as a software dev in the automotive space I would suggest adding the note that this is able to run on POSIX compliant OSes like QNX :) If you would like to chat, hit me up.
46.
by biosboiii 2y ago
I did some research few weeks ago on the topic of database lookup timing side-channels, conclusion is: They don't really exist (for SELECT FROM WHERE commands atleast). https://altayakkus.substack.com/p/timing-side
47.
by biosboiii 2y ago
is the more accurate baseline having food, 3 mortgages, a new phone and a laptop?
48.
by biosboiii 2y ago
Found it out myself, https://binary.ninja/
49.
by biosboiii 2y ago
Does anyone know which decompilation tool produces these graphs, as shown in the blog post?
50.
by biosboiii 2y ago
I am pretty sure that this emergency stop button is a regulatory requirement.
51.
by biosboiii 2y ago
Banning IP Ranges from datacenters is pretty easy and common, surprising they could not combat that.
52.
by biosboiii 2y ago
you can use the official API for that, right? without having to pay ChatGPT and click pixels.
53.
by biosboiii 2y ago
Thing is businesses don't work like side-projects do. Qualcomm is more or less a research company, the main cost of their business is paying engineers to build their modems/SoCs/processors/whatever. They have been workin
54.
by biosboiii 2y ago
nothing specific about next.js but its widely used and this guy chose to provide a easy hands-free tool for that very popular framework
55.
Timing Side-Channel for SQL SELECT WHERE (?)(altayakkus.substack.com)
2 points by biosboiii 2y ago|1 comments
56.
by biosboiii 2y ago
I don't want to clickbait, but I went down a rabbithole of answering my question if the timing of a SQL query, comparing to a string, can be used as a side-channel to more efficiently guess that string. tl;dr is No. Enjoy :)
57.
by biosboiii 2y ago
I work in one of the biggest companies of the world (employee and revenue wise) and it's basically a run-off reaction of well-articulated desk employees jerking each other off that, telling each other that they are so very important. A
58.
by biosboiii 2y ago
Evolution isn't nearly as perfect as you think it may be. For an example, check your feet.
59.
Are neobanks driving the fraud industry? I analyzed 7.594 cases(altayakkus.substack.com)
2 points by biosboiii 2y ago|0 comments
60.
How the EMV protocol beat skimming(altayakkus.substack.com)
2 points by biosboiii 2y ago|2 comments

More ›